PhD candidate · Vision-Language Models

Navendu Brajesh

Building high-signal AI agent systems — where context, not just capability, decides whether an agent works.

Open-source Agent Skills, production patterns, and writing on making LLM agents reliable at scale.

View Work GitHub Medium

Work

Selected open-source projects. Add a new card by appending one object to projects.json.

Loading projects…

Writing

Essays on AI architecture, multimodal models, and platform intelligence.

Read more on Medium →

About

I'm Navendu Brajesh, a PhD candidate researching vision-language models. My work spans multimodal AI, agent systems, and the discipline of context engineering — curating everything an LLM sees so agents stay reliable as sessions grow. My open-source projects distill production patterns into reusable Agent Skills across GitHub Copilot, Cursor, Kiro, Antigravity, and Amazon Q. I write about these ideas on Medium and build them in the open on GitHub.

Connect

For collaboration, speaking, or feedback on open-source work.