AI engineer at U.S. Bank building production LLM systems: a RAG assistant over 500K+ documents for 200+ analysts, LLM eval harnesses (hallucination, PII, prompt-injection), and fine-tuned LLaMA/Mistral models. Outside work I built and published mendrift-mcp (PyPI, listed on the official MCP Registry) an autonomous MLOps incident-response agent with an HMAC-gated approval layer, evaluated on a 19-scenario trajectory harness. Looking for applied AI / ML engineer roles building agentic systems.
suneelkalva1907 · · focus · HN ↗
AI engineer at U.S. Bank building production LLM systems: a RAG assistant over 500K+ documents for 200+ analysts, LLM eval harnesses (hallucination, PII, prompt-injection), and fine-tuned LLaMA/Mistral models. Outside work I built and published mendrift-mcp (PyPI, listed on the official MCP Registry) an autonomous MLOps incident-response agent with an HMAC-gated approval layer, evaluated on a 19-scenario trajectory harness. Looking for applied AI / ML engineer roles building agentic systems.
maxgashkov · · focus · HN ↗