DeepFi Information

DeepFi builds AI monitoring and supervision systems for high-volume, real-time telemetry. The company was founded to take USAD, unsupervised anomaly detection research published at KDD 2020, out of the lab and into production for institutional clients. No labeled failures, adversarial conditions, decisions at stream speed. What we build: * Unsupervised anomaly detection on live data streams: adversarially trained autoencoders running on real telemetry, with no labels required. * Applied LLM and agent systems: an LLM gateway across multiple providers, retrieval-based assistants, and agent workflows, built on the Claude API and open models. * The full path from research to production: modeling, evaluation, deployment, and monitoring, owned end to end, so what ships behaves like what was measured. Our proving ground is decentralized finance: on-chain flows, wallet behavior, and liquidity pools. On-chain markets are open, adversarial, high-volume, and unlabeled by nature, making them the hardest honest test for a detector. The discipline is reliable AI in production; the telemetry happens to be financial. Stack: Python, PyTorch, Claude API, open models, vector search, Docker, cloud.