Reevv
Consultant, Reev Research · Florida, USA, remote
First US client, an idea-stage startup in Florida. Took the problem statement and built the answer from scratch: a Django and Celery pipeline that reads fintech YouTubers, extracts every prediction as a quantifiable, verifiable condition with its stated window, checks it against stock and crypto market data and the web, and scores the creators on their record. Pydantic AI, the OpenAI SDK, LightRAG memory, Logfire monitoring of latency, errors and per-agent spend, YouTube extraction on AWS Lambda. Alongside: HealtheFi (the React Native employee app, the backend and the website), Baddie (the ElizaOS agent runtime rebuilt in Django), and containerised Lambda agents. Ended August 2026.