AI/ML Engineer

Production search and ranking system matching unstructured content to a structured catalogue with 90 percent accuracy. Your Task: 1..Build hybrid search: BM25 plus cosine similarity using sentence-transformers 2. Set scoring thresholds with routing logic for auto-link, review, and discard. Lock boundary test for CI gate. 3. Build feed API: sub-200ms p95 latency, JWT auth, rate limiting, self-healing Stack: Python · TypeScript · sentence-transformers · BM25 · PostgreSQL · Redis · Kubernetes · FastAPI/Express · k6 Deliver: 1. Matching engine with 90 percent accuracy 2. Feed API sub-200ms p95 latency 3. Self-healing plus drift detection 4. CI gate with locked boundary test