Vijay Janapa Reddi 212deffb3a feat(staffml): expand edge/mobile/tinyml tracks with 299 platform-diverse questions ...

Generated 299 new questions across three tracks using gemini-3.1-pro-preview
with platform cycling for vendor diversity:

Edge (99 questions):
  Platforms: Jetson Orin (29), Hailo-8 (55), Coral Edge TPU (34), Qualcomm (25)
  Zones: analyze (48), design (32), diagnosis (19)

Mobile (100 questions):
  Platforms: Apple A17 (26), Snapdragon 8 Gen 3 (23), Tensor G3 (36), Exynos (28)
  Zones: analyze (43), design (32), diagnosis (24)

TinyML (100 questions):
  Platforms: Cortex-M4 (23), ESP32-S3 (34), Cortex-M7+Ethos-U55 (44), nRF5340 (31)
  Zones: analyze (31), design (21), evaluation (17), diagnosis (16)

Track distribution improved:
  cloud: 49.6% (was 52.4%)
  edge: 16.5% (was 15.6%)
  mobile: 14.6% (was 13.5%)
  tinyml: 13.6% (was 12.5%)

19/19 invariant checks pass. Figures regenerated.

2026-04-01 10:29:51 -04:00

..

build_corpus.py

…

expand_tracks.py

feat(staffml): expand edge/mobile/tinyml tracks with 299 platform-diverse questions

2026-04-01 10:29:51 -04:00

extract_taxonomy.py

…

fill_gaps.sh

…

fill_zone_gaps.py

feat(staffml): generate 75 questions to fill ikigai zone gaps

2026-03-31 03:11:12 -04:00

gate.py

…

gemini_fix_errors.py

…

gemini_math_review.py

…

generate_gaps.py

…

generate_hard_questions.py

…

generate.py

…

GUARDRAILS.md

…

preprocess.py

…

review_math.sh

…

run_parallel.sh

…

run_reviews.sh

…

scorecard.py

…

validate_questions.py

…

vault_fill.py

…

vault_invariants.py

…

vault_loop.py

…

WORKFLOW.md

…