Resource Library

All four targets, skills and interview prep separated, every link tagged with how confident I am that it resolves.

Read the verification tags before you click anything. I did not invent a single URL. YouTube video IDs cannot be recalled reliably, and a fabricated link is worse than no link, so entries fall into these tiers: Every entry is a VERIFIED or CANONICAL URL, a verified CHANNEL with a search term, or an explicit NONE. Nothing in this document is a guessed URL.
How to read this document. Four blocks are the plan — 12 (behavioural), 10 (ML system design), 1 (real-time serving) and drill 11.1 (one public repo) — and they run as layers 1, 7, 9 and 10 of 6-month-plan-batches.html, which holds the gates, hours and every purchase decision. Block 3 (graph fraud) is layer 11, conditional on a fraud loop. Every other block is reference: it opens on the condition named in Part 0, not on a week number. Each block header says which. Where this document and the operating plan disagree on what to buy or when, the operating plan wins.

Part 0 — Before the blocks

The plan — gates, hours, purchases — is 6-month-plan-batches.html; this document is the library it links into. Five blocks are in the plan: 12, 10, 1, 11.1 and (conditionally) 3. The other seven open on the condition in the last table of this part, not on a week number.

Application tiering

Quota: two applications a week, logged with date, company, role, level and outcome. Eight in four weeks is part of Batch 0's gate.

TierWhoWhenPurpose
Tier CRecruiters and agenciesHays, Michael Page, specialist ML recruitersNowone of the two per week can be one of theseZero cost to burn. They run practice screens for free and give the fastest read on how the market receives the CV.
Tier BGood companies you would accept but aren't your top fiveMid-size fintech, consultancies with real ML, iGaming operatorsNow, until tier A openstwo per week, loggedLive interview practice with real stakes. If one converts, take it. Every fumbled question goes on the list that reorders the conditional table below.
Tier AFeedzai, Revolut, N26, Nubank, MercadoLibre, Vinted, DatadogOn evidencefive written designs, one delivered verballyThe roles the patents fit. The gate is Batch 1's design gate in the operating plan. Reapplication cooldowns run six to twelve months, so a tier-A loop is spent once.

Diagnostic — twenty minutes, before priming anything

Answer each out loud, timed at two minutes, recorded. If the answer is fluent, skip that primer and go straight to the drill.

SkillSay this out loud in two minutesIf you fumble
DSAMeasured: blank file, DSA templates from memory, twenty minutes. Failed.Layer 3 of the operating plan · the re-take is Batch 0's exit
Control question: how would you detect a fraud ring of forty accounts sharing three devices?Fumbled means articulation, not knowledge. Block 12 first either way.
1Name the four places latency accumulates in a scoring request, and rank them by typical size.Prime 1 — 6 h
2Explain event time versus processing time, and what a watermark does to an event arriving ten minutes late.Prime 2 — 6 h · expected · conditional block
3Why does a forty-account fraud ring evade XGBoost but not a connected-component feature?Prime 3 — 6 h · expected · conditional block
4Name the three components of model risk under SR 26-02 and who owns each.Prime 4 — 5 h · conditional block
5What does pinball loss measure that MASE doesn't, and when does the difference change a decision?Prime 5 — 6 h · conditional block
8Name two retriever metrics and two generator metrics, and the specific failure each one catches.Prime 8 — 5 h · layer 5 of the operating plan, Batch 0
9What is the difference between a resource request and a limit, and which one gets your pod OOMKilled?Prime 9 — 8 h · expected · conditional block; CV fix #3 first
10List the eight sections of a complete ML system design answer, in the order you would deliver them.Prime 10 — 5 h

A tier-A loop before block 10 is done: Prime 10 + Drill 10.1 + Drill 12.1, about 28 hours over two weeks, covers the design and behavioural rounds.

Reference blocks — what opens each

BlockHoursOpens when · reasoning
2 · Streaming & KafkaPrime 2, Drills 2.1, 2.228 hA live fraud loop whose JD lists streaming or real-time features. Genuine gap, and the primer is non-optional when it opens. Until then, the interview answer is “I've built batch and Spark; event-time semantics is the part I'd want to pair on first.”
3.2 · GNN benchmark14 hOnly after Drill 3.1 shows a lift worth the extra complexity. The more interesting result is usually that it doesn't.
4 · Model risk & regulatoryPrime 4, Drills 4.1, 4.222 hA regulated fraud or credit loop reaching a design round. Cheapest credibility in the library, for a room you are not yet in.
5 · Modern forecastingPrime 5, Drills 5.1, 5.230 hA forecasting role reaching a technical round. Stale rather than absent; the primer is a refresh and runs fast.
6 · Retrieval & rankingPrime 6, Drills 6.1, 6.229 hA retrieval-specific role reaching final stages. The two retrieval patents let you discuss this credibly without the drills.
7 · LLM inferencePrime 7, Drills 7.1, 7.226 h~€40A serving-focused role, or month six if still searching. Highest cost, weakest competitive position.
8 · Evaluation drillsDrills 8.1, 8.2 — Prime 8 is layer 5 of the plan22 h~€15A generative role reaching a technical round where the evaluation angle is confirmed. The harness waits for a reason to exist.
9.1 · KubernetesPrime 9, Drill 9.122 hA loop whose JD requires Kubernetes. CV fix #3 (narrow the claim) takes fifteen minutes and removes the liability for every other loop. 9.2 (Terraform, EKS, Airflow) is reference only.
11.2 · Timed SQL and Python15 hThe day a Batch 0 diagnostic test goes red. Free tiers first; a paid platform only through Batch 2 of the operating plan, on a dated failure, one platform not two.

Part A — Skills

1. Real-time model serving & latency Target #1 · Critical

Layer 9 of the operating plan · months 3–4

Your single largest capability gap. Batch scoring is a report; sub-100 ms scoring is a product.

Video

ResourceLinkStatus
MLOps Community channelDeepest free archive on production ML serving anywhereyoutube.com/c/MLOpscommunityCANONICAL
Chip Huyen — Machine Learning is Going Real-Timeapply(conf) April 2021, 30 min. The talk version of the essay in the Written section.tecton.ai/apply/session-video-archive/machine-learning-is-going-real-time/Her other recordings: MLOps Community, search real-time machine learning within the channel.VERIFIED
Ray Summit / Anyscale channelRay Serve, batching, autoscaling for inferenceyoutube.com/@anyscaleCANONICAL
BentoML serving walkthroughsyoutube.com/c/MLOpscommunityThen search within the channel for: BentoMLCHANNEL
Feature store & online/offline skew talksFeast, Tecton; point-in-time correctness is the interview questiontecton.ai/apply/session-video-archive/apply(conf) session archive, nine pages; the fraud-detection workshops are in here. Feature Store Summit recordings: 2024 (Etsy Rivulet, Stripe Shepherd, Chalk sub-5 ms) and 2025.VERIFIED

Written

ResourceLinkStatus
Chip Huyen — Real-time ML: challenges and solutionsStart here. The canonical framing of the whole problem.huyenchip.com/2022/01/02/…VERIFIED
Feature freshness vs model latencyExcellent, contrarian: the model serves in 8 ms on features 40 s oldtacnode.io/post/real-time-ml-inference-feature-freshnessVERIFIED
Snowflake — scaling real-time model servingGPU optimisation, dynamic batching, latency tuningsnowflake.com/en/blog/engineering/scale-real-time-model-serving/VERIFIED
Google Cloud — minimising prediction serving latencyOlder but the architecture reasoning still holdscloud.google.com/solutions/machine-learning/…VERIFIED
Feast documentationBuild one. Do not just read it.docs.feast.devCANONICAL

Prime first

Answer this out loud before touching the drillName the four places latency accumulates in a scoring request, and rank them by typical size.
MaterialWhereHours
Chip Huyen — Real-time ML: challenges and solutionsWritten section, entry 11.5 h
Feature freshness vs model latencyWritten section, entry 20.5 h
Google Cloud — minimising serving latencyWritten section, entry 41.5 h
Snowflake — scaling real-time servingWritten section, entry 30.5 h
Two serving talks from the MLOps Community channelVideo section, entry 12 h
Prime 1b, before Drill 1.2: Feast quickstart plus the Tecton material on point-in-time correctness — roughly 4 h. Do not attempt the parity drill without it; the whole exercise turns on understanding what point-in-time correct actually means.

Drills

How you will be tested“Walk me through serving this model at 10,000 requests per second inside a 50 ms p99 budget. Where does the time actually go?”
Drill 1.1Latency budget teardown~10 h · €0 · local
Problem
Train an XGBoost fraud model, serve it behind FastAPI, then decompose p50/p95/p99 under load into four stages: request deserialisation, feature lookup, model.predict, response serialisation. Then convert to ONNX Runtime and re-measure.
Deliverable
A one-page latency budget table: milliseconds per stage at 100, 500 and 1,000 RPS, native vs ONNX, with the dominant stage identified.
Data & tools
IEEE-CIS Fraud Detection or Sparkov synthetic cards. FastAPI, ONNX Runtime, k6 or locust, py-spy.
Cost
€0 — runs entirely on your M-series.
Stop when
You can state where 80% of your p99 goes without opening the notebook.
Interview claim
“I've profiled a scoring service end to end. Most of the p99 wasn't the model — it was X.”
Failure mode
Reporting the mean instead of p99, and load-testing single-threaded. Both make the numbers meaningless.
Drill 1.2Online/offline parity that fails loudly~14 h · €0 · local Docker
Problem
Define features once in Feast. Generate a point-in-time-correct training set from the offline store and serve the same features online from Redis. Then deliberately introduce skew and prove your test catches it.
Deliverable
A Feast repo (Parquet offline, Redis online in Docker), a point-in-time-correct training set, and a parity test that goes red on injected skew.
Data & tools
Same transaction dataset with event timestamps. Feast, Redis, Docker Compose.
Cost
€0 — all local.
Stop when
Your test fails on the skew you planted, and passes once you fix it.
Interview claim
“Training/serving skew is the first thing I check, and I've built the test that catches it.”
Failure mode
Leaking future information into the training set. That leak is the exact thing the drill exists to teach you to see.

2. Streaming & Kafka feature pipelines Target #1 · High

Reference · opens on a live fraud loop whose JD lists streaming or real-time features

Your Spark background is the shortcut — Structured Streaming is the cheapest entry point you have.

Video

ResourceLinkStatus
Confluent Developer channelThe single best free Kafka curriculum in existenceyoutube.com/@ConfluentCANONICAL
Kafka fundamentals courseTopics, partitions, consumer groups, offsetsApache Kafka 101 — full playlist“Apache Kafka 101” — Confluent, ft. Tim Berglund. 2025 edition. Course text at developer.confluent.io/courses.VERIFIED
Spark Structured Streaming deep diveWatermarks and late data are the interview questionsyoutube.com/@DatabricksThen search within the channel for: structured streaming watermark · Tathagata Das's structured streaming deep dives are the ones to findCHANNEL
Flink Forward channelAwareness only — do not learn Flink nowyoutube.com/@flinkforwardConference channel, playlists per edition (Barcelona 2025 is the latest). Awareness only. Site: flink-forward.org.VERIFIED
Streaming fraud detection architecture talksyoutube.com/@ConfluentThen search within the channel for: fraud detection · Confluent runs financial-services sessions every yearCHANNEL

Written

ResourceLinkStatus
Confluent DeveloperFree structured courses, not marketingdeveloper.confluent.ioCANONICAL
Real-time ML pipelines — ConduktorClean summary of windowed aggregations and the dual-store patternconduktor.io/glossary/real-time-ml-pipelinesVERIFIED
Apache Kafka official docskafka.apache.org/documentationCANONICAL
Spark Structured Streaming guideSplit into sub-pages as of Spark 4.0.0 — use the indexspark.apache.org/docs/latest/streaming/index.htmlCANONICAL
ScyllaDB — feature store latencyConcrete numbers on online store lookupsscylladb.com/2025/07/15/real-time-feature-store/VERIFIED

Prime first

Answer this out loud before touching the drillExplain event time versus processing time, and describe exactly what a watermark does to an event arriving ten minutes late.
MaterialWhereHours
Confluent Developer — Kafka fundamentals courseWritten section, entry 13 h
Spark Structured Streaming guide — event time and watermarking sections onlyWritten section, entry 42 h
Conduktor — real-time ML pipelinesWritten section, entry 21 h
This primer is non-optional. Event-time semantics is not something you will infer from building. Every streaming question in a fraud interview reduces to this distinction, and getting it wrong in a drill means you build the wrong thing confidently.

Drills

How you will be tested“Transaction volume spikes 10× for twenty minutes. What breaks in your streaming feature pipeline, and what do you sacrifice first?”
Drill 2.1Windowed fraud features with honest late-data semantics~14 h · €0 · local Docker
Problem
Replay a transaction dataset as a timed stream. Compute rolling per-card features — count, sum, distinct merchants over 1 h and 24 h — using event time with an explicit watermark. Then inject events 10 minutes late and 2 hours late and show what happens to each.
Deliverable
Docker Compose with Kafka plus a Spark Structured Streaming job writing features to an output topic, and a short note on which late events were absorbed, which were dropped, and why.
Data & tools
IEEE-CIS or PaySim replayed by a producer script. Kafka, Spark Structured Streaming (leans on the PySpark you already have).
Cost
€0 local. Confluent Cloud free tier if you want a managed broker instead.
Stop when
You can explain the watermark trade-off using your own measured numbers.
Interview claim
“I've built windowed streaming features with explicit event-time and late-data handling.”
Failure mode
Using processing time instead of event time. That single choice is what the question is actually testing.
Drill 2.2Break the consumer, then flood it~8 h · €0 · local Docker
Problem
Kill the consumer mid-stream and restart it. Determine empirically what got duplicated, what got lost, and how your offset-commit strategy caused it. Then run the producer at 10× and watch consumer lag.
Deliverable
A two-page write-up: offset/commit timeline, observed duplication on restart, lag curve under flood, and the specific config you'd change in production.
Data & tools
Same stream as Drill 2.1. Kafka CLI tools for lag inspection.
Cost
€0.
Stop when
You can draw the commit timeline from memory and say where duplicates come from.
Interview claim
“I've reasoned about delivery semantics against a real broker, not from a diagram.”
Failure mode
Reading about exactly-once instead of breaking something and observing it.

3. Graph methods for fraud Target #1 · High

Layer 11 of the operating plan · conditional on a live fraud or risk loop (rings, mules, multi-accounting)

Multi-accounting, collusion rings and money mules are graph problems. Your cross-cluster patent is adjacent to this — say so in interviews.

Video

ResourceLinkStatus
Fraud Detection with Graph Neural NetworksPaper-walkthrough format with the references linkedyoutube.com/watch?v=MZGuz-o7Fl0VERIFIED
Stanford CS224W — Machine Learning with GraphsJure Leskovec. The definitive free course. Lectures 1–8 suffice.youtube.com/@stanfordonline → search CS224WCANONICAL
PyTorch Geometric tutorialsantoniolonga.github.io/Pytorch_geometric_tutorials/Longa, Santin & Pellegrini (FBK / Trento). Sixteen tutorials with notebooks; first video JtDgmmQ60x8; code at github.com/AntonioLonga/PytorchGeometricTutorial. Listed on the official PyG resources page.VERIFIED
NVIDIA GTC — GNN fraud sessionsTheir fraud blueprint has accompanying recorded sessionsFraud Detection Using Graph Neural Networks — GTC DC 2025 (DCT51206)Training session: the full workflow from transaction data to a GNN + XGBoost model. Panel version: GNNs for Fraud Detection from Industry Leaders — GTC Spring 2023 (S51704).VERIFIED
Neo4j — graph algorithms for fraudPractical, less academic; good for the "explain it to a stakeholder" anglegraphacademy.neo4j.com/courses/workshop-gds“Graph Data Science in Practice” — free hands-on workshop; the second module is community detection for fraud rings (Louvain, WCC) on a pre-loaded sandbox.VERIFIED

Written

ResourceLinkStatus
safe-graph/graph-fraud-detection-papersCurated, actively maintained. The best single index that exists.github.com/safe-graph/graph-fraud-detection-papersVERIFIED
NVIDIA — GNNs for financial fraudExplains why XGBoost alone misses networked fraud. Useful framing for interviews.developer.nvidia.com/blog/supercharging-fraud-detection…VERIFIED
Graph databases tutorial & surveySection C covers graph anomaly detection specificallyarxiv.org/pdf/2411.09999VERIFIED
GNNs for fraud detection — practical guidefluxforce.ai/blog/graph-neural-networks-for-fraud-detectionVERIFIED
PyGOD — graph outlier detection libraryPython. Build something small with it.github.com/pygod-team/pygodCANONICAL

Prime first

Answer this out loud before touching the drillWhy does a forty-account fraud ring sharing three devices evade XGBoost, but not a connected-component feature?
MaterialWhereHours
Stanford CS224W lectures 1–3 — node embeddings and message passingVideo section, entry 24 h
NVIDIA — GNNs for financial fraudWritten section, entry 21 h
GNNs for fraud detection — practical guideWritten section, entry 41 h
Three lectures, not the full course. You need the intuition for message passing and why neighbourhood aggregation surfaces structure that row-wise models cannot see. Stop there.

Drills

How you will be tested“Forty accounts share three devices and one payout address. How do you find that ring, and why won't your XGBoost model?”
Drill 3.1Quantify what graph structure is worth~12 h · €0 · local
Problem
Engineer graph features over a transaction dataset — shared-device degree, shared-IP degree, connected-component size, two-hop neighbour fraud rate — and measure the lift over a pure tabular XGBoost baseline.
Deliverable
A results table: baseline PR-AUC vs baseline-plus-graph, with feature importances and the single feature that carried the gain.
Data & tools
IEEE-CIS (has device and browser fields) or the Elliptic Bitcoin dataset. NetworkX or igraph, XGBoost.
Cost
€0.
Stop when
You can quote the lift number and name the feature that produced it.
Interview claim
“I've measured what graph structure adds over tabular features on real fraud data — it was X points of PR-AUC.”
Failure mode
Label leakage through the neighbour-fraud-rate feature. Compute it point-in-time only or your result is fiction.
Drill 3.2GNN versus engineered features — publish the loss if it loses~14 h · €0–20 · Colab
Problem
Train a GraphSAGE or GAT node classifier on a labelled fraud graph and benchmark it honestly against Drill 3.1's engineered features on the same split.
Deliverable
A comparison table plus a written conclusion — including the case where the GNN underperforms, which happens often and is the more interesting answer.
Data & tools
Elliptic Bitcoin Dataset, or YelpChi / Amazon fraud graphs via the safe-graph repo. PyTorch Geometric, PyGOD.
Cost
€0 on CPU for Elliptic; Colab free GPU if training drags.
Stop when
You can say under what conditions the GNN's complexity pays for itself.
Interview claim
“I've benchmarked GNNs against engineered graph features and I know when the extra complexity isn't worth it.”
Failure mode
Comparing on different splits or different feature sets. The comparison is the whole point.

4. Model risk, AML & regulatory framing Target #1 · Medium

Reference · opens on a regulated fraud or credit loop reaching a design round

Cheapest credibility in the whole document. Fifteen hours and you speak the language of a compliance-adjacent ML team.

Video

ResourceLinkStatus
ACAMS / AML explainer contentNo reliable videoAML training video is almost entirely vendor marketing or paid ACAMS material. Use the written primary sources instead.NONE
Model risk management SR 11-7 walkthroughsNo reliable videoRead SR 26-02 and SR 11-7 directly — both are short. Explainer videos lag the 2026 revision.NONE
EU AI Act for practitionersartificialintelligenceact.eu/ai-act-explorer/Future of Life Institute, unofficial. The Commission's own tool is now live too: AI Act Service Desk explorer, with a compliance checker — use that one for Drill 8.2, since its output comes with official links.VERIFIED
Feedzai / Featurespace webinarsAlso reconnaissance if you apply therefeedzai.com/resources/Webinars, reports, case studies. Two more for reconnaissance: research.feedzai.com (TimeSHAP, drift tooling) and github.com/feedzai (the Bank Account Fraud dataset suite, a usable substitute in Appendix A).VERIFIED
Explainability for regulated modelsSHAP under audit, adverse action reasoningchristophm.github.io/interpretable-ml-book/Third edition, free online. Read the Shapley values and SHAP chapters; Drill 4.2 is the stability test the book does not run.VERIFIED

Written

ResourceLinkStatus
Federal Reserve SR 11-7The 2011 original. Still the vocabulary everyone uses — but see the row below.federalreserve.gov/boarddocs/srletters/2011/sr1107.pdf PDF — the .htm page blocks some non-US trafficCANONICAL
SR 26-02 — Revised Guidance on Model Risk Management2026 update Interagency: Fed, OCC, FDIC. Rescinds and replaces SR 11-7.federalreserve.gov/supervisionreg/srletters/SR2602.pdf · OCC Bulletin 2026-13VERIFIED
NIST AI Risk Management Frameworknist.gov/itl/ai-risk-management-frameworkCANONICAL
EU AI Act explorerArticle-by-article, searchableartificialintelligenceact.euCANONICAL
EBA guidelines on ML in credit & AMLeba.europa.euCANONICAL
EGBA and IBIA public materialsIf you pursue any iGaming targetegba.eu · ibia.betCANONICAL

Prime first

Answer this out loud before touching the drillName the three components of model risk management and say who owns each in a bank — and note that SR 11-7 was replaced by SR 26-02 in 2026.
MaterialWhereHours
SR 26-02 revised guidance first, then SR 11-7 for the vocabularyWritten section, entries 1–22.5 h
NIST AI Risk Management Framework — core functionsWritten section, entry 21.5 h
EU AI Act explorer — high-risk classification and Annex IIIWritten section, entry 32 h

Drills

How you will be tested“A regulator asks why this specific customer was declined. Walk me through your answer, and then tell me who validated the model.”
Drill 4.1Write the validation document a second line would demand~10 h · €0
Problem
Take the model from Drill 1.1 or 3.1 and write the model validation document an independent risk function would require under SR 11-7 structure.
Deliverable
Six to eight pages: intended use and out-of-scope use, data lineage, assumptions and limitations, performance broken down by segment, stability and sensitivity testing, ongoing monitoring plan, override and fallback procedure, and named revalidation triggers.
Data & tools
Your own model. Fed SR 11-7 letter as the structural template.
Cost
€0.
Stop when
Someone who has never seen your code could mount a credible challenge to the model using only your document.
Interview claim
“I've written model documentation to SR 11-7 structure, including the limitations section.”
Failure mode
Writing a results report instead of a risk document. The limitations and monitoring sections are the deliverable; the accuracy table is not.
Drill 4.2Are your explanations stable enough to give a customer?~6 h · €0
Problem
Generate per-decision SHAP reason codes for declined cases, then perturb inputs slightly and measure how often the top-three reasons change.
Deliverable
A reason-code churn analysis: percentage of decisions where the top reason flips under a 1% feature perturbation, with a recommendation on whether these are fit to disclose.
Data & tools
Your fraud or credit model. SHAP.
Cost
€0.
Stop when
You have a churn number and an opinion on whether it's acceptable.
Interview claim
“I've tested whether model explanations are stable enough to put in front of a customer or a regulator.”
Failure mode
Producing SHAP plots and calling it explainability. Stability under perturbation is the question nobody checks.

5. Modern forecasting Target #2 · High

Reference · opens on a forecasting role reaching a technical round

Prophet, ARIMA and LSTM were the 2019 stack. The gap between what you shipped and what's asked for now is about thirty hours.

Video

ResourceLinkStatus
Temporal Fusion Transformers, EXPLAINEDClearest single explanation of TFT on videoyoutube.com/watch?v=V14qoa5vZ1IVERIFIED
Arvid Kingl — TFT paper clubPaper-reading format; deeper than the aboveyoutube.com/watch?v=M7O4VqRf8s4VERIFIED
Time-series foundation modelsChronos, TimesFM, Moirai — the 2025–26 shiftarxiv.org/abs/2403.07815Ansari et al., TMLR 2024. Code at github.com/amazon-science/chronos-forecasting. Read the benchmark section, not the architecture; the interview question is when zero-shot beats a tuned local model.VERIFIED
Rob Hyndman lecturesAuthor of FPP3; rigorous and freeForecasting: Principles & Practice 3rd ed — playlistHyndman & Athanasopoulos recorded short videos for most sections of the book; they are embedded in otexts.com/fpp3 and collected in this playlist. Background: robjhyndman.com/hyndsight/fpp3_videos.VERIFIED
Conformal prediction for forecastingDistribution-free intervals; increasingly asked aboutarxiv.org/abs/2107.07511Angelopoulos & Bates. The companion repo has the video tutorials and notebooks that run on real model outputs without retraining anything: github.com/aangelopoulos/conformal-prediction.VERIFIED

Written

ResourceLinkStatus
Hyndman & Athanasopoulos — Forecasting: Principles and Practice (3rd ed)Free, complete, canonical. Chapters on hierarchical reconciliation especially.otexts.com/fpp3CANONICAL
TFT original paper (Lim et al.)arxiv.org/abs/1912.09363VERIFIED
Nixtla neuralforecastTFT, N-BEATS, NHITS in one library with consistent APIsnixtlaverse.nixtla.io/neuralforecast/models.tft.htmlVERIFIED
pytorch-forecasting TFT tutorialEnd-to-end demand forecast you can actually runpytorch-forecasting.readthedocs.io/…/stallion.htmlVERIFIED
AI Horizon Forecast — TFT deep divePractitioner newsletter; also covers competition resultsaihorizonforecast.substack.com/p/temporal-fusion-transformer-timeVERIFIED

Prime first

Answer this out loud before touching the drillWhat does pinball loss measure that MASE doesn't, and name a situation where the difference changes the decision.
MaterialWhereHours
FPP3 chapter 5 (evaluation) and chapter 11 (hierarchical reconciliation)Written section, entry 13 h
Temporal Fusion Transformers, EXPLAINEDVideo section, entry 11 h
TFT paper — skim architecture and resultsWritten section, entry 21 h
Nixtla neuralforecast quickstartWritten section, entry 31 h
This one is a refresh, not an introduction. You already own forecasting; only the architectures moved. Expect it to run faster than the six hours budgeted — if it does, bank the time.

Drills

How you will be tested“Prophet gives you a point forecast. Operations needs to decide how much stock to order. What do you actually hand them?”
Drill 5.1Baseline ladder with rolling-origin backtesting~14 h · €0–10 · Colab optional
Problem
Build the full ladder on one dataset — seasonal naive, ETS/ARIMA, LightGBM with lag features, then TFT and N-HiTS — and evaluate all of them with rolling-origin backtesting and quantile loss, not a single split and MAPE.
Deliverable
One table: MASE, pinball loss at q10/q50/q90, and wall-clock training time per model. Plus the point on the ladder where added complexity stopped paying.
Data & tools
M5 Forecasting (retail), Rossmann, or UCI Electricity Load Diagrams. Nixtla statsforecast + neuralforecast, or Darts.
Cost
€0 on CPU; Colab free GPU if TFT is slow.
Stop when
You can name where the ladder stopped paying and defend the number.
Interview claim
“I've backtested classical, gradient-boosted and transformer forecasters on the same data and can tell you where each earns its cost.”
Failure mode
A single train/test split instead of rolling origin. On its own that fails the round, because it means you don't know how forecasts are evaluated.
Drill 5.2Close the loop from forecast to decision~10 h · €0
Problem
Convert the q10/q50/q90 forecast into an ordering decision under asymmetric costs — understock costs four times overstock. Compare the expected cost of deciding from the point forecast against deciding from the quantiles.
Deliverable
An expected-cost comparison and the derived service level, showing which quantile you should actually order at.
Data & tools
Output of Drill 5.1. Newsvendor formulation.
Cost
€0.
Stop when
You can explain why the optimal order quantity is not the median.
Interview claim
“I've closed the loop from a probabilistic forecast to a cost-optimal decision. The quantile you order at isn't the median.”
Failure mode
Optimising forecast accuracy and stopping. Accuracy improvements that don't move the decision are free of value.

6. Modern retrieval & ranking Target #3 · Cheapest high ROI

Reference · opens on a retrieval-specific role reaching final stages

You hold two patents adjacent to this. The gap is two-tower architectures and ANN tuning, not retrieval as a concept.

Video

ResourceLinkStatus
Stanford CS224N / CS224UDense retrieval and neural IR lecturesyoutube.com/@stanfordonlineCANONICAL
Two-tower / dual encoder retrievalp191-covington.pdf (UCSD mirror)Covington, Adams & Sargin, RecSys 2016, DOI 10.1145/2959100.2959190. The canonical two-stage paper. Adrian Colyer's summary is the fastest way in: the morning paper.VERIFIED
James Briggs — vector search seriesPractical FAISS, HNSW, hybrid searchyoutube.com/@jamesbriggsThen search within the channel for: FAISS · his vector-search series is the practical oneCHANNEL
ColBERT and late interactionarxiv.org/abs/2004.12832Khattab & Zaharia, SIGIR 2020. Code and the v2 / PLAID follow-ups at github.com/stanford-futuredata/ColBERT.VERIFIED
RecSys conference — industry trackWhere production retrieval actually gets describedyoutube.com/@acmrecsysOfficial recordings, playlists per year; RecSys 2025 (Prague) talks are linked from recsys.acm.org/recsys25. Industry track only — skip the research track.VERIFIED

Written

ResourceLinkStatus
Eugene Yan's blogBest practitioner writing on retrieval and recsys anywhere. Start with the recsys and search posts.eugeneyan.comCANONICAL
eugeneyan/applied-mlCurated index of how real companies built these systemsgithub.com/eugeneyan/applied-mlCANONICAL
Sentence-Transformers docsBi-encoders, cross-encoders, training with contrastive lossessbert.netCANONICAL
FAISS wikiIndex selection and the recall/latency trade-off, from the sourcegithub.com/facebookresearch/faiss/wikiCANONICAL
MTEB leaderboardKnow what's current before an interview askshuggingface.co/spaces/mteb/leaderboardCANONICAL

Prime first

Answer this out loud before touching the drillReference, not plan. Read this block only if a retrieval-specific role reaches final stages.
MaterialWhereHours
Eugene Yan — retrieval and recsys postsWritten section, entry 13 h
Sentence-Transformers docs — bi-encoders and lossesWritten section, entry 32 h
Your two retrieval patents let you discuss this area credibly without the drills. That is the reason it is cut, not lack of relevance.

Drills

How you will be tested“Your retrieval recall@10 is 60%. What do you change first, and how would you know it worked?”
Drill 6.1Two-tower retriever with hard negatives~14 h · €0–15 · Colab or rented GPU
Problem
Train a bi-encoder with in-batch negatives, measure recall@10 and MRR, then add hard-negative mining and measure the delta. Evaluate in-domain and out-of-domain.
Deliverable
Before/after retrieval metrics with training curves, plus out-of-domain numbers on a BEIR subset to show generalisation.
Data & tools
MS MARCO passage subset, BEIR suite, or Amazon ESCI shopping queries. sentence-transformers, FAISS.
Cost
€0 on Colab free GPU; ~€10 if you rent something faster.
Stop when
You can quote what hard-negative mining was worth in recall points.
Interview claim
“I've trained a two-tower retriever from scratch and measured what hard-negative mining is actually worth.”
Failure mode
Evaluating only on the training distribution. Out-of-domain is where dense retrieval breaks and where interviewers probe.
Drill 6.2Retrieve-then-rank against a latency budget~10 h · €0 · local
Problem
Add a cross-encoder reranker over the top-k candidates from Drill 6.1. Sweep k from 10 to 500 and plot recall@10 against p99 latency.
Deliverable
A Pareto curve with a chosen operating point and a written justification for that point given a stated latency budget.
Data & tools
Output of Drill 6.1. A cross-encoder from sentence-transformers.
Cost
€0 on CPU for a small reranker.
Stop when
You've picked a point on the curve and can defend it against both neighbours.
Interview claim
“I've tuned a two-stage retrieval system against a latency budget rather than chasing NDCG in a vacuum.”
Failure mode
Optimising quality with no cost axis. Every production retrieval decision is a trade, and answering without one reads as academic.

7. LLM inference & serving Target #3 · Real gap

Reference · opens on a serving-focused role, or month six if still searching

You are a consumer of LLMs, not a server of them. This is the difference between "I used LangChain" and "I ran this in production at cost."

Video

ResourceLinkStatus
The Engineering Behind LLM Inference: QuantizationJuly 2026. Memory-bandwidth framing is exactly right.youtube.com/watch?v=1JWnEze9V5gVERIFIED
Optimizing LLM Inference with vLLM and QuantizationDevoxx UK 2026, demo-driven, latency under loadyoutube.com/watch?v=eJBjgakTulgVERIFIED
Umar Jamil — transformer internals from scratchIf KV cache and attention internals aren't second natureyoutube.com/@umarjamilaiCANONICAL
AI Engineer conference channelProduction LLM talks, heavily practitioner-weightedyoutube.com/@aiDotEngineerCANONICAL
PagedAttention / continuous batching explainersarxiv.org/abs/2309.06180Kwon et al., SOSP 2023. This is the vLLM paper; read section 3 for the KV-cache fragmentation argument, which is the whole interview answer.VERIFIED

Written

ResourceLinkStatus
DeepLearning.AI — Fast & Efficient LLM Inference with vLLMShort course, built with Red Hat, June 2026. Best structured entry point.deeplearning.ai/courses/fast-and-efficient-llm-inference-with-vllmVERIFIED
vLLM quantization docsdocs.vllm.ai/en/latest/features/quantization/VERIFIED
vLLM blog — quantization tagFP8, INT4, KV-cache studies with real numbersvllm.ai/blog/tags/quantizationVERIFIED
NVIDIA — post-training quantization guideSmoothQuant, AWQ, AutoQuantizedeveloper.nvidia.com/blog/optimizing-llms-for-performance…VERIFIED
Complete guide to LLM quantization with vLLM — benchmarksNF4 vs AWQ vs GPTQ vs GGUF, with the distinctions made properlyjarvislabs.ai/blog/vllm-quantization-complete-guide-benchmarksVERIFIED

Prime first

Answer this out loud before touching the drillReference, not plan. The only block that costs real money; the weakest competitive position.
MaterialWhereHours
DeepLearning.AI vLLM short courseWritten section, entry 14 h
vLLM quantisation docs and blogWritten section, entries 2–32 h
Revisit at month six if you are still searching, or immediately if a serving-focused role appears.

Drills

How you will be tested“This deployment costs €40k a month to serve. Halve it without users noticing. What do you try first, and how do you prove nothing broke?”
Drill 7.1Find your own quality cliff~12 h · ~€30 · rented GPU
Problem
Serve a 7–8B model with vLLM at FP16, then INT8 and INT4. Measure throughput, time-to-first-token, inter-token latency, and quality — perplexity plus one task-specific eval that matters.
Deliverable
A trade-off table identifying the precision at which quality degrades unacceptably for your chosen task, not in general.
Data & tools
vLLM, LLM Compressor, GuideLLM. Any open 7–8B instruct model.
Cost
~€25–35 for a handful of hours on a rented L40S or A100 (RunPod, Lambda, vast.ai). This is the one drill worth spending real money on.
Stop when
You can quote your own numbers rather than a blog's.
Interview claim
“I've benchmarked quantised serving myself and I know where the quality cliff sits for my workload.”
Failure mode
Measuring throughput at concurrency of one. Single-request numbers tell you nothing about serving economics.
Drill 7.2Size a deployment from a concurrency target~8 h · ~€15 · rented GPU
Problem
Load-test the same deployment at 1, 8, 32 and 128 concurrent requests. Track GPU memory, KV-cache occupancy, and the point where continuous batching stops improving throughput.
Deliverable
A throughput-versus-concurrency curve with the memory ceiling marked and a stated maximum safe concurrency.
Data & tools
vLLM metrics endpoint, GuideLLM or a simple async load generator.
Cost
~€10–15 — run it immediately after Drill 7.1 on the same rented instance and save the spin-up.
Stop when
You can size a deployment from a target concurrency instead of guessing.
Interview claim
“I can size an LLM deployment from a concurrency target and tell you where the KV cache becomes the binding constraint.”
Failure mode
Forgetting to terminate the rented instance. Set a spend cap before you start.

8. LLM evaluation & AI governance Target #3 · Your differentiator

Prime 8: layer 5 of the operating plan, Batch 0 · Drills 8.1 and 8.2: reference, open on a generative role reaching a technical round

You already did this at Oracle. This block is about giving the credential vocabulary and current tooling.

Video

ResourceLinkStatus
LLM Evaluation in Practice: Error Analysis and Reliable Agent TestingApril 2026youtube.com/watch?v=VWrPtb5eWH4VERIFIED
How to Systematically Setup LLM EvalsMetrics, unit tests, LLM-as-a-judgeyoutube.com/watch?v=a3SMraZWNNsVERIFIED
How to Evaluate LLMs: Build a Reliable Evaluation PipelineJune 2026, source code includedyoutube.com/watch?v=rMWudAh5IJsVERIFIED
Evaluating LLM-based chatbots: a reliability frameworkTest cases, regression tests, guardrailsyoutube.com/watch?v=ExA3KSvyzpMVERIFIED
LLM evaluation course playlistyoutube.com/playlist?list=PL9omX6impEuMgDFCK_NleIB0sMzKs2boIVERIFIED

Written

ResourceLinkStatus
freeCodeCamp — AI Evaluation Engineering handbookFull build of a production eval platform. Two weeks old.freecodecamp.org/news/ai-evaluation-engineering…VERIFIED
Arize — LLM evaluation guideLayered evaluator design; LLM-as-judge done properlyarize.com/resources/llm-evaluation/VERIFIED
Evidently — RAG evaluation guideevidentlyai.com/llm-guide/rag-evaluationVERIFIED
Confident AI — LLM evaluation metricsDeepEval author; code examples throughoutconfident-ai.com/blog/llm-evaluation-metrics…VERIFIED
Ragas and DeepEval reposPick one, instrument something you builtgithub.com/explodinggradients/ragas · github.com/confident-ai/deepevalCANONICAL

Prime first

Answer this out loud before touching the drillName two retriever metrics and two generator metrics, and the specific failure each one catches.
MaterialWhereHours
Arize — LLM evaluation guideWritten section, entry 21.5 h
Evidently — RAG evaluation guideWritten section, entry 31 h
LLM Evaluation in Practice: error analysisVideo section, entry 11 h
How to Systematically Setup LLM EvalsVideo section, entry 21.5 h
You already did this work at Oracle across fifteen-plus systems. The primer supplies current vocabulary and tooling, not the concept — treat it as translation.

Drills

How you will be tested“Someone changed a prompt last Tuesday. How do you know whether the system got better or worse?”
Drill 8.1Eval harness built from observed failures, not imagined ones~14 h · ~€15 · API calls
Problem
Build a small RAG system over a corpus you know well. Run it, collect real failures, and derive 60+ test cases from those failures. Separate retriever metrics (context recall and precision) from generator metrics (faithfulness, answer relevancy). Add an LLM judge with a written rubric.
Deliverable
A CI-runnable eval suite that goes red when you deliberately regress the prompt or the chunking strategy.
Data & tools
A corpus in a domain you can judge — Spanish DGOJ gambling regulation, the EU AI Act text, or SEC filings. Ragas or DeepEval; Phoenix or LangSmith for tracing.
Cost
~€10–20 in API calls.
Stop when
Your suite catches a regression you planted without telling it where.
Interview claim
“I've built an eval suite that separates retriever failure from generator failure — which is where most RAG debugging goes wrong.”
Failure mode
Writing test cases from imagination. Cases derived from failures you actually observed are worth ten times as many invented ones.
Drill 8.2Map a real model against EU AI Act obligations~8 h · €0
Problem
Take one model from an earlier drill and work out its classification under the EU AI Act, the documentation obligations that follow, and what a conformity assessment would require.
Deliverable
A gap analysis table: obligation, current state, gap, remediation.
Data & tools
Your own model. The Commission's AI Act Service Desk explorer and compliance checker (linked in block 4), with NIST AI RMF as cross-reference.
Cost
€0.
Stop when
You can classify a system and list its obligations without looking them up.
Interview claim
“I've mapped a production model against AI Act obligations.” Rare, and a direct extension of your Oracle AI safety credential.
Failure mode
Treating it as a legal exercise. Frame it as engineering requirements, which is how a hiring manager will hear it.

9. Kubernetes, IaC & orchestration Target #4 · And a CV liability

Reference · 9.1 opens on a loop whose JD requires Kubernetes; 9.2 is reference only

Kubernetes is on your CV. Until you can defend it, this block isn't optional regardless of which target you choose.

Video

ResourceLinkStatus
TechWorld with NanaBest free Kubernetes and DevOps teaching on the platformyoutube.com/@TechWorldwithNanaCANONICAL
Kubernetes full course for beginnersPods, deployments, services, HPA, resource limitsyoutube.com/watch?v=X48VuDVv0do“Kubernetes Tutorial for Beginners [FULL COURSE in 4 Hours]” — TechWorld with Nana. Watch to the Kubernetes Architecture section minimum (~35 min); the Minikube demo starts after.VERIFIED
Terraform crash courseyoutube.com/watch?v=SLB_c_ayRMofreeCodeCamp, by Sanjeev Thiyagarajan (July 2020) — not a TechWorld with Nana course. Her own Terraform material is on her channel; her full course is paid.VERIFIED
KubeCon / CNCF channelML-on-Kubernetes talks; search within it for Kubeflow and KServeyoutube.com/@cncfCANONICAL
Airflow for ML pipelinesIdempotency and backfills are what get probedacademy.astronomer.io/path/airflow-101Free learning path, Airflow 3, about three hours across twelve modules. The Airflow 2 version is still up if a target runs it.VERIFIED

Written

ResourceLinkStatus
Kubernetes official tutorialsRun them locally with kind or k3d — no cloud billkubernetes.io/docs/tutorials/CANONICAL
AWS EKS WorkshopFree, hands-on, AWS's owneksworkshop.comCANONICAL
HashiCorp Terraform tutorialsdeveloper.hashicorp.com/terraform/tutorialsCANONICAL
Astronomer Academy — AirflowFree and effectively official-gradeacademy.astronomer.ioCANONICAL
KServe / BentoML docsPick one and deploy a model on it end to endkserve.github.io/website · docs.bentoml.comCANONICAL

Prime first

Answer this out loud before touching the drillWhat is the difference between a resource request and a limit, and which one gets your pod OOMKilled?
MaterialWhereHours
TechWorld with Nana — Kubernetes course, first halfVideo section, entry 15 h
kubernetes.io tutorials — run them, don't read themWritten section, entry 13 h
Kubernetes appears on your CV. CV fix #3 — narrow the claim to what you can defend — removes the liability in fifteen minutes for every loop that does not require it. When a loop does, these eight hours convert a liability into a defensible claim faster than anything else here.

Drills

How you will be tested“Your model pod is OOMKilled at three in the morning. Walk me through what you do.” — and note that Kubernetes is currently on your CV.
Drill 9.1Deploy it, break it four ways, diagnose all four~14 h · €0 · local kind/k3d
Problem
Deploy the Drill 1.1 scoring service to a local cluster with resource requests and limits and an HPA. Then deliberately induce four failures: OOMKilled, CrashLoopBackOff, a failing readiness probe, and HPA thrashing. Diagnose each using kubectl alone.
Deliverable
A four-entry runbook: symptom, exact command sequence, root cause, fix.
Data & tools
kind or k3d on your M-series. Remember --platform linux/amd64 when building images.
Cost
€0.
Stop when
You can answer the 3 a.m. question without notes. This is the drill that makes the Kubernetes line on your CV honest.
Interview claim
“I've operated model services on Kubernetes and can debug the four failure modes you actually encounter.”
Failure mode
Deploying successfully and stopping there. Breaking it deliberately is the entire exercise.
Drill 9.2Infrastructure from code, with a working teardown~12 h · ~€20 · AWS
Problem
Terraform a minimal EKS cluster, deploy Airflow onto it, and run a retraining DAG that is idempotent and backfillable. Then destroy everything cleanly.
Deliverable
A Terraform module where destroy leaves nothing behind, plus proof that a 30-day backfill produces identical results to a fresh run.
Data & tools
Terraform, EKS, Airflow (Astronomer Academy for the DAG patterns).
Cost
EKS control plane runs about $0.10/hour. Run it roughly 20 hours across a week and destroy: ~€15–20. Set a billing alarm before your first apply.
Stop when
terraform destroy completes clean and your AWS bill returns to zero.
Interview claim
“I've stood up ML infrastructure from code and torn it down, including backfill semantics.”
Failure mode
Leaving the cluster running. This is the single most common way people blow a €50 budget in three days.

Part B — What actually gets tested

10. ML system design interviews Highest interview weight

Layer 7 of the operating plan · Batch 1

The best-verified block in this document. These are all real mock interviews, not lectures.

Video

ResourceLinkStatus
Full ML Design Mock by ex-Meta Staff EngineerWith feedback afterwards — the feedback is the valuable partyoutube.com/watch?v=9U48NlbOzCUVERIFIED
Instagram — Design a Ranking ModelSenior Meta ML engineer, full mock. Closest to your target #1 loop.youtube.com/watch?v=7_E4wnZGJKoVERIFIED
ETA System for Maps — full mockRegression + real-time constraints; good analogue for fraud scoringyoutube.com/watch?v=l-2m6YXlEjMVERIFIED
Toxic tweet classifier — ML system design mockImbalanced classification under adversarial drift. Nearly your domain.youtube.com/watch?v=ZjNoipQAqRMVERIFIED
Valerii Babushkin — ML System Design InterviewAuthor of the ML system design book; the interviewer's perspectiveyoutube.com/watch?v=0RsmRjar66EVERIFIED
Bonus — recommendation system mocksThree more, same formatqHNlmSnxal0 · jkKAeIx7F8c · 12Ry6kGPQVsVERIFIED

Written

ResourceLinkStatus
Evidently AI — ML system design case collection300+ real company writeups, free. The best single interview-prep asset that exists.evidentlyai.com/ml-system-designCANONICAL
alirezadir/Machine-Learning-InterviewsStructured ML design templates and question banksgithub.com/alirezadir/Machine-Learning-InterviewsCANONICAL
khangich/machine-learning-interviewLong-running, widely usedgithub.com/khangich/machine-learning-interviewCANONICAL
Educative — Grokking the ML InterviewPaid. 6 designed systems, 5 mocks. On the never-buy list in the operating plan (§03) — it duplicates the book plus Drill 10.1. Listed here so you recognise it when a promotion appears.educative.io/courses/grokking-the-machine-learning-interviewVERIFIED
Chip Huyen — Designing Machine Learning SystemsPaid book, and still the one to read cover to coverhuyenchip.com/booksCANONICAL

Prime first

Answer this out loud before touching the drillList the eight sections of a complete ML system design answer, in the order you would deliver them.
MaterialWhereHours
Chip Huyen — Designing ML Systems, chapters 1–2 and the design chaptersWritten section, entry 53 h
Two verified mock interviews, watched actively with notesVideo section, entries 1–22 h
Watch the mocks with a pen and pause after each question the interviewer asks — answer it yourself before hearing the candidate. Passive watching teaches nothing here.

Drills

How you will be testedForty-five minutes, a blank whiteboard, and one sentence of requirements. You are graded on what you ask before you draw anything.
Drill 10.1Five written designs in your target domain~15 h · €0
Problem
Write full system designs for five problems you would actually face: real-time card fraud scoring, bonus-abuse detection on a betting platform, account-takeover detection, AML transaction monitoring, and merchant risk scoring.
Deliverable
Each two to three pages covering requirements and scale, metrics and their business cost, data and labels including label delay and noise, features, model choice, serving architecture with an explicit latency budget, monitoring, and failure modes.
Data & tools
No data needed. Evidently's case collection and Chip Huyen's book as structural references.
Cost
€0.
Stop when
You can deliver any of the five verbally in forty minutes without notes.
Interview claim
This is not a claim — this is the interview.
Failure mode
Designing the model and skipping labels. In fraud, label delay and label noise are half the problem and every interviewer in the domain knows it.
Drill 10.2Read, rebuild, diff~10 h · €0
Problem
Take ten case studies from the Evidently collection in fraud, risk or recommendations. Read only the problem statement, design it yourself in twenty minutes, then read what the company actually did and record the difference.
Deliverable
A diff log: for each case, what you missed and why you missed it.
Data & tools
evidentlyai.com/ml-system-design.
Cost
€0.
Stop when
Your diffs shrink to details rather than whole missing components.
Interview claim
Builds the pattern library that separates a fluent design round from a stumbling one.
Failure mode
Reading the answer first. The twenty minutes of your own struggle is where the learning is.

11. Coding, SQL & ML fundamentals under time pressure Underrated risk

Drill 11.1: layer 10 of the operating plan, months 3–4 · Drill 11.2: layer 4, opens the day a Batch 0 diagnostic test goes red

Nine years of seniority does not exempt you from a timed coding screen, and rustiness here fails more senior candidates than any design round.

Video

ResourceLinkStatus
Emma DingRenamed from “Data Interview Pro”. DS/ML interview breakdowns, SQL and stats.youtube.com/channel/UCAWsBMQY4KSuOuGODki-l7ACANONICAL
StatQuestFor refreshing the fundamentals you'll be asked to explain simplyyoutube.com/@statquestCANONICAL
Exponent channelSource of most of the verified mocks aboveyoutube.com/@tryexponentCANONICAL
Advanced SQL for analyticsWindow functions, QUALIFY, sessionization, funnelsEmma Ding channelThen search within the channel for: SQL interview · she covers window functions and the classic funnel questionsCHANNEL
Live Python coding for ML rolesyoutube.com/@tryexponentThen search within the channel for: machine learning coding interviewCHANNEL

Written

ResourceLinkStatus
DataLemurSQL and DS questions from real companies, not LeetCode-style puzzlesdatalemur.comCANONICAL
StrataScratchstratascratch.comCANONICAL
NeetCodeIf a DSA screen appears, do the blind-75 onlyneetcode.ioCANONICAL
Architecture Patterns with PythonFree online. Directly addresses your "research code isn't rigorous" answer.cosmicpython.comCANONICAL
interviewing.ioPaid anonymous mocks with real engineers. Worth more than any book here — and it is Batch 3 of the operating plan, behind an entry gate, because it is the one spend wasted if you arrive empty-handed. Exponent's free mocks cover Batch 0.interviewing.ioCANONICAL

Prime first

Answer this out loud before touching the drillWhat does a property-based test give you that thirty assertions on mock data don't?
MaterialWhereHours
Architecture Patterns with Python, chapters 1–3Written section, entry 43 h
ruff, uv and mypy --strict docsAstral and mypy documentation1 h

Drills

How you will be testedA timed screen, usually 45 minutes, usually SQL plus one Python problem. Seniority does not exempt you and rustiness fails more senior candidates here than in any design round.
Drill 11.1Take one drill repo to production standard~20 h · €0
Problem
Choose your strongest drill above and add strict typing (mypy --strict), tests including one property-based test on a transform, data contracts with pandera, ruff and pre-commit, and CI that goes green.
Deliverable
A repository that passes strict type checking, has meaningful coverage, and installs cleanly from a fresh clone.
Data & tools
Whichever drill you pick. Architecture Patterns with Python (free at cosmicpython.com).
Cost
€0.
Stop when
A stranger can clone it and run the tests without asking you anything.
Interview claim
Answers “maintainable, tested code and code reviews” directly — and answers your own admission that your research code isn't rigorous.
Failure mode
Adding tests that assert nothing. One property-based test on a real transform is worth thirty assertions on mock data.
Drill 11.2Timed sets, not leisurely ones~15 h · €0–30
Problem
Thirty SQL problems weighted toward window functions, cohorts, funnels and sessionisation, each on a fifteen-minute timer. Plus twenty Python problems on a thirty-minute timer.
Deliverable
A log of every problem that exceeded the timer, with the reason — concept gap, syntax rust, or misread question.
Data & tools
DataLemur or StrataScratch (both partly paid); NeetCode blind-75 if a DSA screen appears.
Cost
€0 on free tiers. A paid platform only through Batch 2 of the operating plan, on a dated failed test, and one of the two, never both.
Stop when
Your timer-exceeded rate is under 10%.
Interview claim
None. This one is purely defensive.
Failure mode
Practising untimed. The failure mode in a real screen is pressure, not knowledge, and untimed practice doesn't train for it.

12. Staff-level behavioural & influence Your stated weak point

Layer 1 of the operating plan · six stories this week, the rest in Batch 1

You told me you had no evidence of influencing engineering strategy and then produced a CV that says otherwise. This block is about retrieval, not learning.

Video

ResourceLinkStatus
Staff engineer archetypes and scopenoidea.dog/glueTanya Reilly. Full transcript with slides; the talk recording is LeadDev New York 2019. The written version is faster.VERIFIED
LeadDev channelConference talks on senior IC influence and technical strategyyoutube.com/@LeadDevCANONICAL
STAR method for senior engineersyoutube.com/@LeadDevThen search within the channel for: staff engineer · conference talks on senior IC scope, not generic STAR adviceCHANNEL
How I Prepared for ML System Design Interviews at MetaCovers the behavioural half of the loop tooyoutube.com/watch?v=XN2ymraj27gVERIFIED
Negotiation for senior offersRelevant given your €90K floorhaseebq.com/my-ten-rules-for-negotiating-a-job-offer“Ten Rules for Negotiating a Job Offer” — Haseeb Qureshi. Note the slug differs from the title — that is why searching the title alone failed. Part 2 is How Not to Bomb Your Offer Negotiation.VERIFIED

Written

ResourceLinkStatus
StaffEng — story archiveReal staff engineers describing how they got there. Read five, find your archetype.staffeng.comCANONICAL
Will Larson — Irrational ExuberanceHis blog; the free precursor to the bookslethain.comCANONICAL
Tanya Reilly — The Staff Engineer's PathPaid. The one that maps exactly to your stated gap.noidea.dogCANONICAL
ADR templatesWrite two about real Oracle decisions. That's your evidence artifact.github.com/joelparkerhenderson/architecture-decision-recordCANONICAL
levels.fyi negotiation guidelevels.fyiCANONICAL

Prime first

Answer this out loud before touching the drillNo primer. This block is retrieval of your own history, not acquisition of anything new.
MaterialWhereHours
StaffEng story archive — read five stories for calibrationWritten section, entry 11.5 h
Read the five stories only to calibrate what staff-level scope sounds like when someone else describes it. Then close the tab and write your own.

Drills

How you will be tested“Tell me about a time you changed a senior stakeholder's mind.” You told me you had no evidence of this. Your CV says otherwise. This block resolves that.
Drill 12.1Ten stories, each with a number and a conflict~8 h · €0
Problem
Write ten STAR stories from your real history with mandatory coverage: three where you changed a senior person's mind, two where you were wrong and caught it yourself, two on mentoring, two on ambiguous problems you scoped, one on a production failure you owned.
Deliverable
Ten stories, each containing one hard number, each deliverable in two minutes.
Data & tools
Your own history. StaffEng archive for calibration on what staff-level scope sounds like.
Cost
€0.
Stop when
You can tell any of the ten cold, in order, without hesitating.
Interview claim
Directly closes the gap you named in our conversation.
Failure mode
Stories with no number and no conflict. “We collaborated well” is not a story; it's a mood.
Drill 12.2Two retroactive architecture decision records~6 h · €0
Problem
Document two real technical decisions from your Oracle work in ADR format: context, options considered, decision, consequences, and what you would do differently now.
Deliverable
Two ADRs, each one page.
Data & tools
ADR templates from joelparkerhenderson/architecture-decision-record.
Cost
€0.
Stop when
Both read as if written at the time of the decision, not afterwards.
Interview claim
This is your concrete evidence artifact for the technical-influence gap. Bring them to the interview.
Failure mode
Writing the decision without the options you rejected. The rejected options are what demonstrate judgement.
If a link is dead. Every entry keeps its exact title and creator in the left column or the qualifier, so a dead VERIFIED link is recoverable by searching title and creator together — YouTube titles are stable even when IDs get re-uploaded, and arXiv IDs never change. Papers: search the title on arxiv.org directly.

Appendix A — Dataset index

Everything the drills need. Several are shared across drills, so download once.

DatasetWhat it isLicence / accessUsed by
IEEE-CIS Fraud DetectionKaggle competition data~590k e-commerce transactions with device, browser and card fields. The default fraud dataset.Kaggle account; competition rules, non-commercialDrills 1.1, 1.2, 2.1, 3.1, 4.1
Sparkov / synthetic credit card transactionsFully synthetic card transactions with merchant and geo fields. No licence friction.Kaggle; permissive, syntheticSubstitute anywhere IEEE-CIS is used
Bank Account Fraud (BAF) suiteFeedzai Research, NeurIPS 2022 datasets trackSix synthetic account-opening fraud tables built to be biased, imbalanced and temporally dynamic, with demographic fields. Designed for fairness and drift evaluation, which makes it the better choice for Drills 4.1 and 4.2 — and it is reconnaissance on a tier-A target.Via github.com/feedzai; check the licence on the repo before anything publicSubstitute for IEEE-CIS in 1.1, 3.1, 4.1, 4.2
PaySimSimulated mobile-money transfers with labelled fraud. Naturally streaming-shaped.Kaggle; open, syntheticDrills 2.1, 2.2
Elliptic Bitcoin Dataset~200k nodes, 234k edges, labelled licit/illicit. A genuine graph, not a table you graphed.Kaggle; research useDrills 3.1, 3.2
YelpChi / Amazon fraud graphsStandard benchmarks in the graph-fraud literature. Comparable numbers exist in papers.Via safe-graph GitHub; research useDrill 3.2
M5 Forecasting — AccuracyWalmart hierarchical retail demand. Intermittent, hierarchical, realistic.Kaggle; competition rulesDrills 5.1, 5.2
UCI Electricity Load Diagrams370 clients, 15-minute resolution. Clean, no account needed.UCI ML Repository; openDrill 5.1 (alternative to M5)
MS MARCO passage rankingLarge-scale passage retrieval with relevance judgements.Microsoft; research licenceDrill 6.1
BEIR benchmark suiteEighteen retrieval datasets for out-of-domain evaluation. This is where dense retrieval breaks.GitHub; mixed per-datasetDrill 6.1 (evaluation only)
Amazon ESCI shopping queriesProduct search relevance, multilingual including Spanish.GitHub, Amazon Science; permissiveDrills 6.1, 6.2
EU AI Act text · Spanish DGOJ regulation · SEC filingsDense regulatory corpora in domains you can judge — which is what makes them good RAG test beds.PublicDrills 8.1, 8.2
Pick one fraud dataset and stay on it. Drills 1.1 through 4.2 chain together — the model you train in 1.1 is the model you serve in 1.2, break in 9.1, and document in 4.1. Switching datasets between them wastes the compounding. IEEE-CIS is the better choice if you want realism; Sparkov if you want zero licence friction.

Appendix B — What this actually costs

Against your €50/month ceiling. Nine of the twelve skill blocks cost nothing. This table prices infrastructure for drills; platform and book purchases are decided in 6-month-plan-batches.html §02–03 and nowhere else.

DrillResourceCostFree path
7.1 + 7.2LLM quantisation & concurrencyRented L40S or A100 — RunPod, Lambda, vast.ai~€40$0.80–2.00/hr, ~20–25 hrsColab Pro (€10/mo) fits a 7B at INT4 but not a clean FP16 baseline. Run both drills back-to-back on one instance.
9.2Terraform + EKS + AirflowEKS control plane + small node group~€20~$0.10/hr control planeSubstitute ECS Fargate or run Airflow locally in Docker. The Terraform skill transfers either way.
8.1RAG eval harnessLLM API calls for generation and judging~€15Use a small local model via Ollama for generation; keep the paid model as judge only.
6.1Two-tower trainingGPU hours~€10Colab free tier or Kaggle notebooks (30 GPU-hours/week). Genuinely sufficient here.
11.2SQL and Python setsDataLemur or StrataScratch€0–55/yrBatch 2 only, on a dated failed testFree tiers first, plus the LeetCode SQL and pandas study plans. The purchase decision lives in the operating plan, not here.
All othersNine of twelve blocksYour M-series, Docker, kind€0
Two spending rules before you touch a cloud console. Total across all twelve blocks: roughly €85–115, which fits your ceiling if you spread the paid drills over two or three months and keep the free path where it's offered. For the four-block plan the infrastructure cost is €0: blocks 12, 10, 1 and 11.1 all run on your M-series. The first euro appears only when a conditional block opens (8.1's API calls, or a rented GPU for 7.1).

Appendix C — On keeping all of this private

I built these as private drills because you asked for that. I'd make exactly one exception.

Your case for private is real: portfolio repos take three to five times as long as drills, the polish is time not spent applying, and a mediocre public repo is worse than nothing. I'm not arguing with that for eleven of the twelve blocks.

Drill 11.1 is the exception. You have a six-month employment gap, a CV claiming Kubernetes and AWS you told me you haven't shipped, and no public artifact anywhere. One public repository does three jobs simultaneously: it substantiates the “maintainable, tested code and code reviews” line that appears in the JD, it gives the gap a name you can say out loud, and it satisfies the “open-source contributions” nice-to-have in both Super ML reqs. It costs about twenty hours because you're building the repo regardless — the only marginal cost is a README and making it public.

Private drills produce capability with zero external evidence. A six-month gap is precisely the situation where evidence is the scarce thing, not capability.