Berkeley EECS Honors · Systems / Infrastructure Engineer · Class of 2027
Incoming intern on Databricks' Traffic Platform — networking, security, and performance engineering for large-scale distributed infrastructure in Go and Rust.
Previously Amazon SDE Intern — built real-time IoT platform processing 10K+ events/min across 100+ sensors with <500ms latency on ECS Fargate.
I build systems projects around LLM agent security, low-latency C++ infrastructure, and multi-cloud developer tooling.
Selected experience and projects. Each item links to code, benchmarks, or details.
Security proxy for tool-using LLM agents that detects prompt-injection attempts before high-risk tool calls. MCP is the protocol agents use to call external tools; LLMFirewall sits between agent and tools enforcing taint-aware trust levels, capability gating, heuristic detectors, and lightweight semantic-drift scoring.
Exchange-style matching engine with deterministic price-time priority, FIFO fairness, and reproducible benchmark harnesses. Optimized with slab allocators, cache-conscious order layout, branch reduction, and CPU pinning. Includes Binance US market data connector, TAQ replay, and crash recovery.
Synthetic local benchmark — hardware spec and harness in repo README.
Desktop deployment tool that generates provider-specific serverless configs, provisions infrastructure with Pulumi, streams real-time deploy logs, and normalizes deployment workflows across Cloudflare Workers, AWS Lambda@Edge, and Vercel. Monaco editor, per-provider ENV injection, and deployment history.
2025
Amazon
SDE Intern
TypeScript · AWS Lambda · ECS Fargate · DynamoDB · SQS · Timestream
2026 (incoming)
Databricks
SWE Intern — Traffic Platform incoming
Incoming SWE intern on the Traffic Platform team — networking, security, and performance engineering for large-scale distributed infrastructure.
Go · Rust · Distributed Systems · Networking