I build practical AI systems: agent workflows, RAG backends, ML prototypes, healthcare automation, and systems-level tooling.
I am building Agent_Sudo as a solo project: an authorization, delegation, provenance, and verifiable-audit engine for AI agents.
The goal is simple: agents can act, but risky actions need scoped authority, source-aware policy checks, and an audit trail you can verify after the run.
Proof points: GitHub · Latest release v0.5.5 · PyPI · MCP Registry · Glama · Dev.to writeup
|
Multi-agent workflows, tool orchestration, retrieval pipelines, evaluation loops, and automation systems that move beyond chat-only interfaces. |
Computer vision, burnout detection, healthcare AI experiments, model fine-tuning, and pragmatic ML prototypes with usable interfaces. |
Backend services, MLOps pipelines, Android ROM/device work, cloud deployment, observability, and durable engineering workflows. |
|
|
|
|
|
|
|
|
- Building a Permission Gateway for MCP Agents: What I Learned After Letting AI Run Local Tools
- From Chatbots to Action-Bots: The Great Shift to “Agentic Workflows” in AI
- How I Was Built: The Architecture of Bujji, an AI Companion
- From RAG to REFRAG: Building Trustworthy AI in Healthcare
- This AI Reads Your Face and Tells You If You’re Burned Out Built in 3 Days
