They Do Not Survive Contact With Production
Running 28 autonomous agents across four companies, with self-healing, governance, and cost attribution.
Speaking
I do not pitch with slides. I pitch with telemetry.
I speak on production multi-agent systems, AI in regulated industries, local-first inference, and what it takes to build a company that actually runs on AI. Every talk is grounded in a real operation you can verify, not a demo.
The credential
A machine-verified snapshot of the operation behind the talks, live from one workstation.
ARCHIVEone workstation
as of 2026-07-020
services running
0
scheduled timers
0
agents in fleet
0
commits
0
hourly snapshots
0d
days uptime
Talk topics
Nine angles, adaptable to your track.
Trust Requires Constraints
Putting AI inside a HIPAA-regulated billing operation, where what an agent cannot do matters more than what it can.
Self-Improving Agents With Receipts
Eval-gated overnight training with frontier-lab statistical rigor, at the marginal cost of electricity.
397 Billion Parameters on a MacBook
Streaming a Mixture-of-Experts model from NVMe in hand-written C and Metal, and the 38% speedup from deleting the cache.
I Do Not Pitch With Slides, I Pitch With Telemetry
Proving a one-person, always-on AI company with services, timers, commits, and snapshots you could audit.
Four Industries, One Operator
What it takes to hold equity in four AI businesses at once and run them with an agent fleet.
Deterministic Before Generative
An anti-hallucination pattern for regulated AI: facts by exact lookup, the model only judges and writes.
From Theology to the Server Room
What a graduate degree in language and theology taught a builder about memory, dependency, and autonomous systems.
A Bot That Joins Your Google Meet
Full-duplex voice agents where no bot API exists, running tools live by voice mid-call.
Speaker bios
Copy-ready, by audience.
Technical program
Nabbil Khan builds and operates production multi-agent systems and hand-writes inference engines that run frontier open-source models on consumer hardware. He runs a 28-agent fleet across four companies and can back every claim with machine-verified telemetry.
Business / keynote
Nabbil Khan holds equity in four businesses and runs them with a fleet of autonomous AI agents he built. A former classical scholar and Bitcoin miner, he is the founder of Tensor Solutions, Mentor Agile, and MedicalBiller.AI.
Healthcare program
Nabbil Khan is building the first AI-native medical billing operation, from a HIPAA-audited voice agent that calls payers to a denial-management system backed by a 4.7-million-row knowledge base. He focuses on trustworthy AI in regulated workflows.
Longer bios, headshots, and one-sheets on request · official@nabbilkhan.com