About me
I got into AI the practical way: by trying to make it do real work, and paying close attention to where it fell apart.
At Tesla I spent two years as a data engineering intern. I built data pipelines and quality tools, and helped engineers use AI in their daily work. At Ford, as a product manager intern on the software engineering side, I built internal AI tools and reusable workflows, and led AI events across teams. Now I’m the technical lead at Trexoros, where I build software and AI workflows. That includes the product, the backend, the agent tools and the tests. AI agents draft content for several brands, and people decide what ships.
Along the way I’ve built the other things on this site: forecasting models, creative AI tools, a robot arm controlled from a browser, and a 3D machine explainer. The domains differ, but the pattern stays the same: I get close to the real problem, build the smallest thing that could work, and measure whether it did.
Most people skip that last step, and it is the one I care about most. A demo shows that something can work. A measurement shows how often it does.
I’m finishing a computer science degree at Arizona State in December 2026, and I’m open to consulting. My favorite clients have an AI idea, or an AI system that almost works, and want someone to make it dependable.
- Based in
- Austin, Texas
- Now
- Technical Lead, Trexoros
- Studying
- B.S. Computer Science, ASU · Dec 2026
- Focus
- Agents · MCP · Evaluation
- Building with AI
- Since 2024
- Available for
- Consulting
Experience
Technical Lead
I lead software and AI workflow delivery end to end: the product interface, backend, agent tools, integrations, testing and deployment. Agents draft campaigns across several brands through an MCP server I rebuilt and benchmarked. People keep approval and publishing. Read the case study.
Product Manager Intern · Software Engineering
I built internal AI tools and reusable workflows, collaborated across teams, and organized a 130 person hackathon.
Data Engineer Intern
Two years of data work on the Cybertruck and Model Y: building data pipelines and data-quality tools, and helping engineers put practical AI workflows to use.
Cyber Security Intern
Cyber governance.
Data Scientist Intern
ETL pipelines and visualization with Python and Power BI.
Software Engineer Intern
JavaScript and QuickBase work, with RMM vendors and AWS.
Robotics and Coding Teacher
Part time. I taught robotics and programming logic.
Cyber Security Specialist Intern
Red team and blue team penetration testing, in the cloud and on premises.
Information Technology Intern
Support tickets, onboarding and infrastructure management.
B.S. Computer Science
The theory under the practice: algorithms, systems and software engineering. My capstone contribution was calendar sync and notifications for a multi-team platform.
“He ramped quickly, worked independently, and knew when to pull in help.”
How I work with AI
I direct the work, read what the agent did, correct it and keep going. These are three real long sessions from my own projects. Each bar is one slice of the session. Its height is how many tools the agent called in that slice.
Metadata only: counts of tool calls in 32 equal time slices. No conversation text or tool output is shared. The recorded span includes breaks, so it is not working hours, and a call count is not a productivity score.
Daily AI usage
Building with it every day is how I learned where it helps, where it struggles, and what still needs a person.
Daily usage · Dec 25, 2025 to Sep 29, 2026
Source: my ccusage export, captured September 29, 2026. This is partial history, not everything since 2024. A blank day means no entry in this export, not necessarily no AI use. No cost figures are shown.
Work with me
Audits, builds and evaluations for agents, MCP servers and AI workflows. It helps to include three things: what the workflow does today, what “working” should mean, and any constraints I should know about.
The calendar loads from Cal.com when you open it. Nothing loads from Cal.com before that.