alvis
65218762be
Sets up the full Phase 0 foundation: - pnpm workspaces + turbo build graph; native module build approval - packages/shared-types: HTTP contracts (Tip, Auth, Integrations, User) - services/api: Express modular monolith with better-sqlite3/drizzle - auth: Google OAuth2 + PKCE via openid-client v6, cookie sessions - integrations: Todoist OAuth2 connect/disconnect, token vault - recommender: RandomPolicy over Todoist tasks, feedback sink - user: profile, consent capture, full account deletion (GDPR) - apps/web: Next.js 15, three pages (sign-in → connect → tip) - tip page: black canvas, hold-to-act gesture, action sheet - PWA manifest + theme - ml/serving: FastAPI stub implementing the POST /score contract - infra: docker-compose (core/full profiles), Dockerfiles, CI skeleton - .env.example with all required vars documented Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
ml/
Python. Owns models, features, training, online scoring.
| Dir | Role | Phase |
|---|---|---|
serving/ |
FastAPI online scorer (/score), called by recommender |
1 |
features/ |
feature definitions + store adapter (Feast later) | 1 |
pipelines/ |
batch feature + training DAGs (Prefect/Airflow) | 4 |
registry/ |
MLflow-backed model registry integration | 4 |
experiments/ |
A/B assignment + multi-armed bandit policies | 4 |
notebooks/ |
research; never imported by production code | — |
Principles
- Every model has a model card in
registry/describing inputs, offline metrics, fairness checks, and rollout history. - Online inference must be stateless and < 50ms p99.
- Training reads from the offline feature store; serving reads from the online feature store; definitions are shared (no train/serve skew).
- Shadow deploys before any policy change that affects real users.