Structure messy workflows
Turn ambiguous processes into clear states, inputs, interfaces, and review points.
Work
Technical writeups and builds
This section is ready for project writeups. The first entries will focus on AI-assisted content workflows and LLM-enriched data pipelines.
Add a Markdown file in src/content/work to publish one.
How I build
Practical systems around ambiguous workflows
Use models as components
Build around LLMs and tools where they improve text-heavy or judgment-heavy work.
Keep validation close
Design workflows with checks, evals, review, and failure handling near the output.
Ship practical interfaces
Care about the product layer: what the user sees, controls, trusts, and repeats.
About
From data science to AI engineering
My background is in applied data science: ML/NLP workflows, data pipelines, model-driven analysis, and systems built around imperfect real-world data. I am now focused on the engineering layer around AI: useful interfaces, reliable workflows, tool use, validation, and deployment.