📝 Assignments: Phase 6 (Scale, Polish, Interview)
W23 - Scale & Cost
🎯 A23
- Load: 1,000 concurrent runs + 500 concurrent streams (mock providers with realistic latency)
- Profile → fix the top 3 bottlenecks → re-test; a scaling journal with graphs
- Partition
run_events; archive completed runs to S3 (claim check); measure the DB size reduction - $/run by use case at 1M runs/month; cut 40% with routing/caching while holding eval scores
- “100× plan” design doc: cells, a sharded engine, a multi-region control plane
🧠 Cognitive: Constraint flip: the same system at ½ budget. What do you cut? Score: __/28
W24 - Portfolio & Interview Transfer
🎯 A24
- README: the problem, architecture, the 12 hard problems, numbers, demo video, how to run it
- 6 blog posts published; an ADR index
- Mock system design: “Design an AI workflow platform” and “Design an LLM gateway” (45 min each, recorded)
- Mock LLD: “Design the workflow engine classes” in 90 min
- 10 STAR stories drafted from Orbit (Behavioral & STAR)