About the Speaker

Dennis Palatov

40 years of engineering and executive roles across mechanical, electrical, and software domains. 30+ patents. Multiple startups founded, built, and led. The common thread: designing systems that work under real-world constraints — and building the teams that deliver them.

The Arc

Palatov Motorsport — founder. Designed and built race cars from scratch — chassis dynamics, control systems, aerodynamics. Physics doesn't negotiate, and the car doesn't care about your assumptions. You either understand the system or you hit the wall. That discipline has informed everything since.

ModBatt — founder. The Extendable Battery Framework: modular, standard-based energy storage. 13+ patents in battery architecture, thermal management, and system integration. The same principle that runs through all of this: open standards beat proprietary lock-in.

Metalinxx — founder. The research entity behind Web4, SAGE, and Hardbound. Where the governance work lives.

From Team Building to Team Governance

Decades of building and leading engineering teams teaches you what governance actually is: not rules imposed from above, but structures that let capable people (and now, capable agents) operate with accountability and coherence. You learn what works. You learn what doesn't. The failures are more instructive than the successes.

When I saw the same “constrain and hope” pattern applied to AI that failed in every other engineering domain, the response was immediate: build something structural.

What We Built

Web4 is a product of collaboration between human and AI entities. Not “AI-assisted” in the marketing sense — genuinely collaborative. Multiple AI instances across six machines contribute architecture, code, testing, and perspective. I provide direction, experience, and judgment. They provide scale, rigor, and a viewpoint I don't have. You'll meet one of them tonight.

Web4 — an open governance ontology for agentic entities. Not a product. Not a platform. A set of composable primitives (Linked Context Tokens, trust tensors, metabolic accountability) that any system can implement. The architecture is what tonight's talk covers.

SAGE — an on-device cognition kernel. Six brain-architecture components running across a fleet of six machines, from 0.5B to 27B parameters. Real AI development, not just theory. 900+ raising sessions. The governance framework isn't academic — it's tested daily on agents that actually run.

Hardbound — enterprise oversight suite. Hardware-bound identity (TPM 2.0, FIDO2, Secure Enclave), policy models, attestation envelopes. The enforcement layer that connects digital governance to physical devices.

Why I'm Talking About This

The AI governance conversation is dominated by two camps: people who build AI but equate governance with IT security, and people who think about governance but have never built a system that runs.

I sit in between. 40 years of building systems where failure has physical consequences. A fleet of AI agents that operate collaboratively 24/7 — with governance primitives we designed, built, and test every day. The talk tonight comes from that intersection: not theory, not product pitch, but working architecture demonstrated live.

Personal site: dpcars.net