About

I build AI systems end to end — from retrieval and agent orchestration to evaluation, observability, tool interfaces, and developer workflows. My recent work focuses on production-minded LLM applications, multimodal agents, MCP infrastructure, and resource-efficient language models.

I like taking ambiguous problems, turning them into working systems, and then making the systems easier to inspect, evaluate, and improve.

Open Source

Independent Open-Source Contributor — Agenta AI

Contributed fixes and improvements to Agenta, an open-source platform for building agents and automations.

Course Reviewer — Towards AI Academy, Agent Engineering

Reviewed Agent Engineering course material, suggesting code fixes and content improvements with a focus on practical, production-oriented agent engineering.

Current Focus

Efficient language-model inference

Research

Exploring architectural modifications for resource-constrained deployment, including attention redesign and knowledge distillation.

On-device AI

Early research

Investigating small language models for mobile deployment, including next-word prediction on Android hardware.

Agent systems & tooling

Active

Building frameworks, MCP servers, and observability infrastructure for production agent workflows.

Links