Location: Mexico City, Mexico Remote: Yes — US timezone Willing to relocate: Yes Technologies: Python, FastAPI, LangGraph, LangChain, RAG, pgvector, PostgreSQL, Redis, Celery, PyTorch ExecuTorch, LoRA fine-tuning, Next.js, React Native, TypeScript, Docker Résumé/CV: https://docs.google.com/document/d/1zYNrg8UpUF1jrW3TFBJ3FST1... Email: diego@diegocm.dev

22, self-taught. Won a $100K Meta Llama Impact Grant for Bluai and presented it at LlamaCon on 2025

Bluai (https://play.google.com/store/apps/details?id=com.bluai.app&...) is a hurricane early-warning app live on Google Play in Mexico, built offline-first because during a hurricane the network is the first thing to go. An online RAG pipeline and an on-device Llama 3.2 (ExecuTorch, NPU-accelerated) sit behind one interface, picking 1B or 3B based on device RAM, so the UI never knows which brain answered. When everything is down it falls back to Bluetooth/Wi-Fi mesh chat over a multi-hop wire protocol.

Email me one paragraph about something broken or missing in your product. I'll build a working prototype and send you the repo plus a short video of it running, within 48 hours. Free. No call, no interview, no invoice. If you like it, we talk. If not, you keep the code.

Flexible on compensation for the right team. I'm optimizing for who I'd be learning from, more than thinking in numbers.