Interview Prep Manual
From LeetCode to Lithography
Semiconductor data science interviews don't test what LeetCode tests. They test whether you can apply algorithmic fundamentals to physical systems where experiments cost $50,000 per wafer, ground truth arrives 60 days late, and a 200-millisecond latency miss destroys $2.5 million of silicon.
Part 0 is free. Parts 1 through 5 require full access.
How this pairs with missions
25 of the platform's 70 missions directly target the concepts interviewers test. Work through this manual first, then use the tagged missions as active practice. The remaining 46 missions cover your first year on the job, waiting for you after you get the offer.
Note types used throughout
DS Translation
How this fab concept maps to a DS pattern you already know.
Bootcamp Trap
The standard ML answer that fails in a fab context.
Why It Matters
The financial or safety consequence of getting this wrong.
Good News
Where your existing skills transfer directly.
00
The ResetFree
What you must unlearn
Three constraints that reshape every algorithmic decision, the 11 banned techniques that instantly fail interviews, the six numbers that signal credibility, and the vocabulary you must own cold.
The Interview Reality CheckThe Banned ListThe Numbers to MemorizeThe Vocabulary Checkpoint
01
The Eight Archetypes
Every interview draws from these eight failure modes
Eight recurring failure modes from real fabs, each with a 60-second story, a DS translation, a structured answer framework, and an explicit LeetCode parallel. Master these and you can answer virtually any behavioral or case study question.
The Averaged ArcThe Boiling FrogThe Daylight Saving DisasterThe Ghost ExcursionThe OFAT ExtinctionThe Frozen ThermocoupleThe Invisible Reticle KillerThe Life Safety Breach
02
The Technical Foundation
Working code for live interviews
The five SQL join patterns, Python data engineering patterns, the SPC decision tree, Nyquist-Shannon for sensor data, and the ISA-95 hierarchy. Every code block is something you could write cold in a shared editor.
SQL for Fab DataPython for Fab DataStatistical Process ControlSampling and Signal ProcessingThe ISA-95 Hierarchy
03
The ML Arsenal
Algorithms that run in production
Virtual Metrology architecture with mandatory OOD detection, FDC implementation hierarchy, model selection under latency and interpretability constraints, the edge deployment pipeline, and walk-forward validation patterns.
Virtual MetrologyFault Detection and ClassificationModel Selection Decision TreeEdge DeploymentValidation Methods
04
The Interview Simulation
Practice every format you will encounter
SQL live-coding with three practice problems, Python coding challenge with a full endpoint detector implementation, STAR-format behavioral stories from the archetypes, hostile review defense scripts, system design whiteboard framework, resume translation, and the questions you ask them.
SQL Live-CodingCoding ChallengeBehavioral RoundHostile ReviewSystem DesignYour Resume TranslatedQuestions You Ask Them
05
Company Deep Dives
Tailored prep for your target company
ASML (EUV stochastics, overlay control), Lam Research (endpoint detection, chamber matching), Applied Materials (cross-step pattern detection), Intel/TSMC/Samsung (fleet monitoring, yield at scale), and startups. Specific vocabulary, technical challenges, question banks, and the interview day checklist.
ASMLLam ResearchApplied MaterialsIntel / TSMC / SamsungSmaller Companies and StartupsThe Interview Day
Companion: Field Manual
After you get the job, the Field Manual covers your first year: seven layers from foundational constraints to production algorithms. It makes your interview prep look easy by comparison.
Open Field Manual Unlock all six parts
Full access to the Interview Prep Manual, Field Manual, 70 missions, and 5 capstones.
Unlock full access, $149