Software Engineer @ Microsoft
Agentic AI platforms · LLM orchestration & evaluation · distributed systems

I build agentic AI systems — agent runtimes, tool-calling, multi-agent orchestration, and LLM evaluation — on top of a distributed-systems foundation (7+ years, from datacenter control planes to agent platforms).

• Agentic AI infrastructure — agent runtimes, tool-calling / function calling, multi-agent orchestration, sandboxed execution
• LLM evaluation — rubric-based scoring, LLM-as-judge / validator-agent patterns, mid-execution validation
• Multi-provider orchestration — cost-aware routing across Anthropic Claude, OpenAI, Azure OpenAI, Ollama, with MCP
• Distributed systems — scheduling, checkpoint/restore, consistency, and fault tolerance at scale

Distributed systems research (Virginia Tech, Systems Lab):

• Ekko: Fully Decentralized Scheduling for Serverless Edge Computing — IPDPS '25 · IEEE
• Capybara: An Edge-Friendly Distributed Object Store for Diverse Serverless Functions — Middleware '25 · ACM
• Toward an Edge-Friendly Distributed Object Store for Serverless Functions — APSys '24 · ACM

Part of the Amplifier agentic-AI ecosystem:

• amplifier-agent — agentic AI CLI + engine library + JSON-envelope wire protocol (TypeScript & Python SDKs)
• amplifier-sdk — typed SDK over the Amplifier kernel
• amplifier-grove — Amplifier modules for Grove IDE integration