Vijay Janapa Reddi ef7ad0a0b4 feat(interviews): fill 68 gap questions + fix 7 math errors from R2 ...

Gap fill (68 new questions):
- Cloud: 17 questions (L1/L2 architecture, data, reliability, optimization)
- Mobile: 24 questions (L1-L6+ across data, networking, reliability)
- Edge: 8 questions (L4-L6+ networking, parallelism)
- TinyML: 19 questions (L1-L6+ data, optimization, networking)
- All with verified napkin math, MCQ for L1-L3

Math review Round 2 (Opus, per-track deep review):
- Cloud: 817 reviewed, 3 errors (H100/A100 TFLOPS swaps, INT4→INT8)
- Edge: 798 reviewed, 0 errors
- Mobile: 768 reviewed, 2 errors (per-minute→per-hour, 48→45 TOPS)
- TinyML: 712 reviewed, 2 errors (unit conversion cascade, throughput)
- Combined R2 error rate: 0.22% (7/3095)

Corpus: 3,248 total (3,161 published + 87 archived)

2026-03-24 08:49:46 -04:00

..

core

…

perf

…

__init__.py

feat(interviews): fill 68 gap questions + fix 7 math errors from R2

2026-03-24 08:49:46 -04:00