Cornell CS / ex-Cofounder & CTO @ Drip (YC P26) / systems, HPC, AI infra
I build infrastructure across the stack: agent tooling, memory systems, cloud backends, and the low-level pieces that make software feel fast. Cornell CS, ex-YC founder, currently going deep on distributed systems and AI infrastructure.
current loop
read the paper -> build it from scratch -> stress it -> measure -> internalize
| | | | |
systems + primitives only, failure latency / real
networking no shortcuts modes throughput intuition- cloud development intern at Computime via the HKSTP Global Internship Program (Hong Kong) — multi-tenant IoT cloud backends, AWS/Azure, telemetry pipelines, and compliance automation at scale
- building systems depth from the primitives up: networking, distributed systems, and concurrency in Go — implemented from scratch to build real intuition
- previously Cofounder & CTO at Drip (YC P26), 2026
- studying CS + Math at Cornell, with a bias toward systems, HPC, AI infrastructure, and tools that make teams faster
| area | what I like about it |
|---|---|
| cloud & systems infra | multi-tenant backends, telemetry pipelines, IaC, observability, and security/compliance automation |
| distributed systems | queues, idempotency, durability, backpressure, observability, and failure modes |
| AI tooling | agents with useful memory, local-first workflows, execution logs, and retrieval surfaces |
| performance | C/C++, Rust, CUDA, vectorized Python, and the parts of HPC that turn theory into throughput |
| product engineering | sharp interfaces, boring reliability, fast iteration, and fewer layers between problem and fix |
| project | shape |
|---|---|
| meepcall | local-first macOS meeting/call recorder with ScreenCaptureKit, whisper.cpp, and optional Recall/Anthropic add-ons |
| meeple | local semantic-memory store: ingest personal data, build a markdown wiki, and query it with an LLM |
| millennium-data-quality-25-26 | factor research and backtesting toolkit with IC analysis, walk-forward validation, HRP, stress tests, and tearsheets |
| dl-ec | interactive visualization of RoPE and rotary position embeddings |
| geometry-dash-rl | teacher-student RL pipeline with a Geode shared-memory bridge |
- site: lucas-309.github.io
- email: lucashe309@gmail.com
- linkedin: linkedin.com/in/lucashe309



