Question answering over financial filings (10-K/10-Q, investor decks) with cited answers: hybrid dense+BM25 retrieval, reranking, chart extraction, and an AST-sandboxed calculator. FastAPI, CLI, Helm.
Terminal score: 0–1000 raw, weighted across 4 dimensions. Public score: 0–10 normalized (shown in the 30-day stars chart above).