Experience
Member of Technical Staff Intern
Tensormesh
March 2026 - Present
Foster City, CA
- Tensormesh, the team behind LMCache, helps enterprises cut GPU costs by offloading KV caches to external storage for reusability during inference.
- LMCache is used by teams at Nvidia, Google, Redis, AWS, Red Hat, Cohere, and more.
Open Source Contributor
Meta
October 2025 - October 2025
Remote
- Contributed to Pyrefly, Meta's high-performance Python type checker and language server written in Rust.
- Refactored the error summarization subsystem by enforcing standard naming conventions across module definitions and import paths.
Software Engineer Intern
Quantum Rings
June 2025 - August 2025
Chicago, IL
- Delivered 19 PRs, 43 contributions, and 15 completed GitHub issues across the internship, adding ~15K LOC and removing ~3.6K LOC while reviewing code and driving schema refactors.
- Migrated execution data from the user entity to a dedicated relational table with FKs, modularizing schema and ensuring test suite stability with no performance regression.
- Implemented a telemetry aggregation background worker (AWS SQS + TypeORM) to asynchronously roll up user execution activity, improving scalability and simplifying downstream analytics queries.
- Designed and deployed queue-driven execution processing to decouple heavy telemetry operations from the API, reducing request latency and enabling horizontal scaling.
- Built full-stack admin analytics dashboards with NestJS, Next.js, and Recharts, integrating SQL time-bucket aggregation and timezone-safe filtering to track user growth, active usage, and execution volume.
Projects
Hostess
Docker Compose for Production. Declare your whole stack — frontends, APIs, databases, caches, and workers — in one hostess.yml, run hostess deploy, and your entire app ships to managed Kubernetes with service discovery, managed databases, TLS, preview environments, and a Studio dashboard built in. No infrastructure to wire up, no YAML sprawl to maintain.
Redis Operator
A Kubernetes operator for Redis, inspired by CloudNativePG. Supports standalone, sentinel, and cluster modes with direct pod and PVC management. Implements a fencing-first failover strategy to prevent split-brain scenarios and ensure data consistency during node failures.
AutoMoE
A modular, multi-task self-driving system based on a Mixture-of-Experts (MoE) architecture. Uses specialized neural networks for specific perception tasks and a gating network to combine their outputs, aiming for a more modular, interpretable, and efficient model for navigating complex environments.
Launchpad
AI-powered student–startup matching platform. Uses pgvector for semantic search across student profiles and startup listings, with async BullMQ scoring pipelines that rank candidates by fit. Helps early-stage startups find talent and students discover opportunities aligned with their skills.
Matchbox
An AI-powered research matching platform that connects university students with labs using semantic search and LLM-based fit scoring. Streamlines the application process with a centralized portal, replacing fragmented email workflows with automated candidate discovery, resume parsing, and compatibility ranking.
Education
Computer Science
Bachelor of Science
University of Chicago
September 2024 - June 2028
Relevant Courses:
- Introduction to Computer Science I
- Introduction to Computer Science II
- Systems Programming I
- Systems Programming II
- Mathematical Foundations of Machine Learning
- Discrete Mathematics
- Theory of Algorithms
- Machine Learning
- Foundations of Distributed Systems
Physics
Bachelor of Arts
University of Chicago
September 2024 - June 2028
Relevant Courses:
- Mechanics
- Electricity and Magnetism
- Waves, Optics, and Heat
- Intermediate Mechanics
- Quantum Mechanics I
- Quantum Mechanics II