AI systems with a clear business purpose.

I work at the application and systems layer: deciding where AI can be useful, how it fits into existing operations, and what architecture will make it dependable enough to use.

The work begins with a practical question. What should improve for the customer or the organization, and how will we know whether it has?

What I focus on

Application architecture

Connecting models, data, interfaces, and existing systems into a coherent product. The architecture needs to support the workflow, the people using it, and the team responsible for operating it.

Retrieval and grounded answers

Designing how a system finds the information it needs, respects access boundaries, and makes its sources available for review. Good retrieval starts with understanding the material and the questions people need to ask of it.

Agents and tool use

Defining what an agent can do, which tools it can use, and when a person should make the decision. Multiple agents can help when responsibilities are genuinely separable; the coordination needs to justify its complexity.

Private and local AI

Working through deployment choices for sensitive data and internal workflows. The right approach depends on control, performance, operating cost, and the organization’s ability to maintain the system.

AI, creative work, and intellectual property

Exploring how content, rights, and provenance can support useful commercial applications. At Amplist, this connects technical design with the decisions creators and rights holders need to make about their work.

How I approach the work

1. Define the useful outcome. Identify the user, the workflow, and the business reason to change it.
2. Understand the constraints. Examine the data, integrations, rights, operating environment, and failure consequences.
3. Build a narrow working version. Test the hardest assumptions before expanding the scope.
4. Evaluate the system. Check useful output, failure behavior, cost, latency, and the need for human review.
5. Prepare for operation. Define ownership, monitoring, change control, and the path from prototype to ongoing use.

A longer history of applied intelligence

My interest in applied AI predates today’s models. I founded Style du Jour in 2006 to explore personalized wardrobe and shopping experiences. That work produced a prototype and two issued patents in algorithmic fashion matching.

Later technology advisory work included a private AI and data cluster for a growing biotech company. Today, my focus is on application architecture and commercially useful AI systems.

Current work at Amplist

Amplist helps creators and rights holders evaluate AI opportunities and develop products around their content, rights, and ideas. My role connects the technical architecture with the product and business decisions required to move forward.

Explore Amplist · Discuss a technology leadership opportunity