About Ken Tabor

Bio

I deliver working software to solve customer problems.

I'm a product-minded software engineer with an eye toward design. I'm drawn to ambitious ideas, but I tend to approach them from first principles: understand the problem, surface the trade-offs, and connect what's technically possible with what actually works.

I've spent my career moving across engineering, product, UX, APIs, developer experience, and now AI platforms. Somewhere along the way I discovered that my best work tends to sit at the intersection of building systems and helping people make sense of them.

Along the way I worked with product teams building TripCase, partnered with enterprise architects to improve API design standards and developer experience, advised executives exploring AI-powered platform strategy, and helped teams turn emerging technology into things customers could actually use.

As generative AI accelerated, it became the technical problem I couldn't ignore. I helped take Sabre's MCP Server from zero to one, trained teams experimenting with generative AI, participated in product discovery, represented the customer in technical decisions, and served as lead author of the company's technical white paper on AI agents.

I have a bias toward action and enjoy getting hands-on. Sometimes that means writing code. Sometimes it means sketching an architecture, testing a product idea, explaining a new technology, or asking the annoying-but-useful question: What problem are we actually trying to solve?

What I'm Working On

Right now I'm in a self-directed R&D chapter focused on one deceptively simple question:

What does it take to build world-class AI agents?

I'm studying that question by building things. My work spans generative AI, agent engineering, developer experience, product strategy, tools, memory, context, evaluation, and the control surfaces humans need to work effectively with increasingly autonomous software.

I'm especially interested in the point where AI systems stop being impressive demos and start becoming useful: software that can see what needs to be done, do meaningful work, check itself, and remain understandable and controllable by the people depending on it.

I share much of what I learn through writing, prototypes, talks, and conversations with other builders. Teaching is a useful forcing function. If I can't explain an idea clearly, I probably don't understand it well enough yet.

Code!

I still like making things. Current tools and technologies in my orbit include:

• OpenAI models and SDKs, agent frameworks, MCP, local models, Hugging Face, and the wider generative AI ecosystem
• Python, JavaScript, Node.js, React, HTML, CSS, Liquid, Pandas, and plenty of command-line tools
• AI-assisted software engineering with tools such as Codex
• AWS including Lambda, S3, CloudFront, Route 53, API Gateway, DynamoDB, CloudWatch, Certificate Manager, and Cognito
• Git, VS Code, IntelliJ, and whatever else helps turn an idea into working software

The specific tools change quickly. The durable part is the practice: understand the system, build something, test it, learn from reality, and improve it.

Sharing What I Learn