Portfolio

I build AI products at the boundary between models and real software: provider integrations, evaluation systems, production services, developer examples, and the infrastructure that makes fast inference useful.

Selected Projects

Developer ecosystem · Cerebras

Fast inference where developers already work

Built and maintained integrations across LangChain, Pydantic AI, Flowise, Dify, Cline, Kilo Code, and OpenCode.

Outcome: native Cerebras support, model discovery, reasoning controls, and production-oriented examples across major AI tooling ecosystems.

See LangChain releases

AI product platform · Cordial

From first service to evaluated product capability

Built the company's first production Python AI service and connected it to the core PHP platform under a tight evaluation deadline.

Outcome: the resulting AI product capabilities earned a 5 out of 5 score in the 2024 Forrester Wave evaluation.

Enterprise retrieval · Corteva

Conversational access to a large data catalog

Helped build a GPT-4 retrieval system across roughly 200 enterprise data sources.

Outcome: a knowledge interface designed to serve 21,000 employees.

AI automation · Cordial

Generation, agents, voice, and optimization

Developed customer-facing capabilities for message generation, audience building, agent automation, outbound voice workflows, and send-time optimization.

Outcome: send-time optimization increased clicks by 30%.

Open-Source Engineering

Other Product Work

Scholarly Contributions

Thought Leadership Pieces

Conference Engagements

HuggingFace Models

Proud contributor of 36 unique models to the open-source NLP library, HuggingFace.

University Teaching

Over the years, I’ve had the pleasure of transferring knowledge to over 840 students in 21 different university classes.