I work on things that need to run, not just compile.
Most of my work focuses on AI models, training and inference pipelines, and the tooling required to evaluate, debug, and operate them in real environments.
When existing tools don’t fit, I build small, focused solutions to close the gap.
I don’t strictly separate research, engineering, and infrastructure — if something blocks the work, I dig into it and fix it.
- AI model training, inference, and evaluation
- Performance and reliability issues in AI systems
- Tooling and infrastructure to support experiments and deployments
- Debugging unclear, slow, or partially working systems
 |
 |
 |
 |
 |
 |
|---|---|---|---|---|---|
| Python | Node.js | Git | Linux | Shell | Terminal |
 |
 |
 |
 |
|---|---|---|---|
| Kubernetes | Cloud | Database | iptables |
These tools are not goals by themselves — they’re things I use when the work requires them.
I’ve done software security research in the past.
Today, that experience shows up as a habit of testing assumptions, understanding failure modes, and not trusting systems blindly.
I occasionally write about things I’m working on or learning:
https://blog.zonetwelve.io
(Updates are irregular and usually tied to real work.)
Last updated: 2026