AI Engineer • AI Researcher • Open-Source Contributor • Systems Hacker • Startup-Minded Builder
I'm an AI software engineer focused on shipping real-world systems that blend high performance with rapid iteration. I’ve worked on LLM-integrated backends, prompt orchestration protocols, memory management systems, and AI-powered internal tools—plus contributed to open-source projects at GitLab.
💼 I currently lead AI infra + backend engineering at Moonnox, a VC-backed startup building enterprise LLM agents.
At Moonnox, I:
- 🧠 Designed and deployed stateless LLM APIs on AWS Lambda & FastAPI
- 🧮 Built a Supabase-backed multi-turn memory engine to improve agent context retention
- 🔗 Created a Model Context Protocol (MCP) to sync prompt memory across tools like Salesforce, Notion, Jira, and SharePoint
- 📉 Implemented latency monitoring + error tracing with Datadog + PostHog
- 🚀 Overhauled CI/CD pipelines across GCP to reduce deploy time from 30m → <5m
- 📊 Built internal scoring tools to track agent hallucination, drift, and task completion
I work closely with our CTO and frontend team to ship fast, debug hard problems, and iterate directly with real users.
On the side, I contribute to StreamGale, a Rust-based AI infra project for:
- ⚙️ Low-latency inference at the edge
- 🛰️ MQTT-based real-time streaming pipelines
- 🧠 Lightweight federated learning across devices
In 2024, I also won a national AI + data science hackathon — featured in Times Square and 1871 for competitive programming and LLM work.
- 🥇 Winner, AI & Data Science Hackathon and National Competitive Programming Competition (featured in Times Square + 1871)
- 🧠 Moonnox: Built AI orchestration layer, context memory system, and prompt integration pipelines used in live enterprise workflows
- 🔧 StreamGale (side project): Rust-based low-latency edge inference infra
- 📦 Marvel Platform: Maintained open-source tools for collaborative research workflows
- 🌍 GitLab Contributor: Contributed to GraphQL testing and RuboCop pipelines
Thanks for visiting 🙌