Build the judgment
behind the answer.
A focused engineering workbench for system design, coding patterns, engineering field notes, and production lessons.
JUDGMENT
NOTES
THE PERSON BEHIND SYSTEM CRAFT LAB
I build platforms and write about the engineering decisions behind them.
My work spans storage, streaming, logistics, and advertising systems. System Craft Lab brings that perspective to system design, AI fundamentals, and practical engineering guides.
Previously at Coupang · DoorDash · eBay · Microsoft · Amazon
Read my field notesIndependent publication. Views are my own.
Choose your learning vector
Move through a track or cross-link ideas as your interview loop demands.
System Design
Architecture, trade-offs, distributed systems, and the decisions behind them.
AI Foundations
Build understanding through worked examples, small Python experiments, and exercises, from fundamentals to models and agents.
Engineering Field Notes
Production lessons, technical deep dives, and decisions from real engineering work.
Engineering Digest
A daily signal feed distilled from high-quality engineering blogs.
Computing for Science
Python pipelines, numerical experiments, and scientific measurements from public data.
Coding Patterns
Reusable problem-solving patterns, algorithms, and interview-ready practice.
Turn reading into a plan
System Design Fundamentals
Build a repeatable interview framework: requirements, estimates, architecture, bottlenecks, and trade-offs.
5 steps · 80 min →PATH 02Senior Interview
Combine coding mechanics, scalable architecture, operational judgment, and concise interview communication.
5 steps · 140 min →PATH 03Staff / Principal
Practice the judgment signals senior interviewers look for: boundaries, failure containment, cost, and evolution.
5 steps · 106 min →PATH 04Distributed Systems
Trace one large system through delivery semantics, partitioning, storage tiers, query fan-out, and recovery.
5 steps · 95 min →PATH 052-Week Sprint
A focused fourteen-session plan that alternates architecture, coding, drills, and current engineering signals.
14 steps · 395 min →High-signal reading
Selected articles on engineering decisions and scientific measurement.
Most Frequently Co-Purchased Items
Find each product's most frequent shopping-basket companion with exact pair counts, deterministic ties, and tested Java and Python solutions.
NAT: Addresses, Ports, and Connection State
Trace packet translation and return-path state, calculate hot-destination port capacity, and design NAT egress around failure boundaries, observability, and explicit access policy.
Training, Validation, and Test Sets
Give fitting, model selection, and final evaluation separate data roles, then reproduce a small experiment without allowing test results to influence the chosen procedure.
Continue exploring
Most Frequently Co-Purchased Items
Find each product's most frequent shopping-basket companion with exact pair counts, deterministic ties, and tested Java and Python solutions.
NAT: Addresses, Ports, and Connection State
Trace packet translation and return-path state, calculate hot-destination port capacity, and design NAT egress around failure boundaries, observability, and explicit access policy.
Training, Validation, and Test Sets
Give fitting, model selection, and final evaluation separate data roles, then reproduce a small experiment without allowing test results to influence the chosen procedure.
Case Study: From a Vague Request to a Verifiable AI Task
Turn a troubleshooting-assistant proposal into a bounded document recommendation task, with explicit evidence, a baseline, permissions, and testable acceptance criteria.
Business Metrics and Model Metrics
Connect prediction quality to service outcomes, compare errors under explicit costs and capacity limits, and measure whether a model improves the work it supports.
When You Don't Need Machine Learning
Recognize tasks that are better served by rules, queries, search, optimization, or human review, and compare the evidence and operating costs before adding a model.