Miguel Ingram

I'm the founder of Black Box Research Labs, a forensic technical due diligence consultancy focused on AI-augmented codebases. My forensic engine performs polyglot static analysis (Python, JavaScript, Go) inside air-gapped containers, applies the AIV (AI Integrity Verification) protocol for cryptographically-anchored audit evidence, and produces board-ready briefings.

I've completed forensic audits of five AI agent frameworks (CrewAI, Continue, HumanLayer, LangChain, Agno) producing 44 verified findings, data-integrity validation for the Oklahoma Conservation Commission's Blue Thumb citizen-science monitoring program, and an audit for Workspring (founder Jacob Askey provided a public testimonial). I'm a published AI researcher (arXiv:2512.07109, 302 fine-tuned Transformer models), recently presented peer-reviewed research at the OCLWA 2026 conference, and a US Navy veteran (Nuclear Electronics Technician).


Session

06-13
15:20
45min
Who Reviews the Reviewers? A Live Forensic Audit of an AI Agent Framework
Miguel Ingram

A CTO defined six autonomous AI agents and merged the PR eleven seconds later. Zero reviews. This isn't anecdote: it's structural. I built a forensic engine to audit AI-augmented codebases, and I'm running it live on stage on a popular AI agent framework. Air-gapped static analysis, cryptographic provenance, per-blank human attestation, and five questions every defender should be able to answer about their own infrastructure Monday morning.

Track 1: In the Future (Emerging Tech, AI and Automation, Seeing Round Corners)
Track 1 I.T.F. (Mengler Conf. Rm.) (2nd Floor)