About
Hi! I'm Collin, and I like building things end to end. I'm a Computer Science and Engineering student at Santa Clara University with a minor in Artificial Intelligence, and I spend most of my time on the parts of a product other people would rather not touch: the network layer, the storage layer, and everything holding the interface together.
I just finished up an internship at Rockliffe Systems, where I engineered WebRTC calling, a SQLite storage layer, and more into an Electron XMPP client called AstraChat. I have also been working as a Full Stack Developer at Sport Endorse since June 2025, where I architected and shipped a headless Next.js platform serving 10,000+ athletes, brands, and agencies. I also work as a Computer Networks Research Assistant at SCU, building RL simulations for integrated sensing and communication 6G systems.
In my spare time you can find me competing in hackathons, weighlifting, hiking around California, or playing basketball. I'm always happy to connect!
Experience
Software Engineer Intern · Rockliffe Systems
Engineered WebRTC audio and video calling in an Electron XMPP application with dynamic STUN/TURN discovery, multi-protocol ICE (IPv4/v6/TCP), device switching, and call recovery. Migrated storage from LocalForage to a local SQLite cache with MAM history and cross-device sync for reactions, receipts, and roster state, and hardened the network layer by resolving XML stanza disconnects and optimizing Tigase server SQL queries.
Full Stack Developer · Sport Endorse
Architected and deployed a headless full stack platform into production serving 10,000+ global athletes, brands, and agencies. Executed Blue/Green zero-downtime deployments and drove a 50% increase in leads with a 100% Lighthouse score through regional A/B testing, localized caching, and technical SEO. Work across a multi-timezone Agile team spanning the US, Europe, Dubai, and South Africa.
Computer Networks Research Assistant · Santa Clara University
Build DDQN ISAC simulations in PyTorch using NVIDIA Sionna digital twin models and an ns-3 network testbed. Researching PASS NOMA wireless systems, RSMA, and semantic communication.
B.S. Computer Science and Engineering, Artificial Intelligence Minor · Santa Clara University
Coursework in data structures and algorithms, operating systems, compilers, databases, networks, web development, and AI. Member of ACM, the AI Collaborative, IEEE, and BroncoSec, and a regular at AWS x INRIX, Hack for Humanity, and NVIDIA hackathons.
Projects
Stash — Workspace Snapshot Manager
Cross-platform desktop app built with Tauri that saves and restores window layouts, VS Code paths, terminal working directories, and browser tab state. Wrote Win32 API bindings to read process memory directly, and paired a browser extension over a local token-authenticated HTTP server to rebuild full tab state. Windows and Linux installers ship from a GitHub Actions pipeline.
LA Rental Home Finder
AI-driven real estate comparison tool using AWS Bedrock and Llama 3 to score deal quality and detect price gouging across 300+ listings, built to find honest rentals in Los Angeles after the fires.
AskRON
The Reliable Option for Navigation — a Chrome extension for AI assisted browsing, powered by Claude on AWS Bedrock with a Flask backend. Built for the AWS x INRIX hackathon.
Budget Travel Planner
Travel planner that searches bus, rail, and mixed-mode routes through the Transitland API and compares them on price, duration, CO2 emissions, and intermediate stops. City search autocompletes against Nominatim, the itinerary is held in Zustand state, and a finished itinerary exports to CSV.
Flitter
Full-stack social platform with user creation, posting, following, profile editing, and password authentication, persisted to MongoDB through Prisma.
Amtrak vs. Plane Trip Calculator
Parses flight tickets from major airlines through the SkyScanner API and prices Amtrak routes alongside them, comparing cost, departure and arrival times, and station locations.
Research
Phyiscal PASS NOMA Testbeds for 6G Networks
In Progress!
Mobility Aware Dual Waveguide ISAC Pinching Antenna NOMA for Resilient Mission Critical Networks
IEEE Open Journal of the Communications Society
A mobility aware resource allocation framework for pinching antenna systems on a dual waveguide architecture, where one waveguide tracks mobile users by radar while a second serves them over NOMA. We derive the sensing channel from first principles, establish a tracking accuracy constraint that holds with roughly 30 dB of margin, and solve the joint grouping, assignment, power, and repositioning problem with a double deep Q network. Evaluated on disaster response and stadium emergency scenarios against exhaustive search, a cost matched greedy heuristic, and published sum rate optimal power allocation.
Channel Disparity Engineering for NOMA Leveraging Pinching Antenna Systems
Santa Clara University
Successive interference cancellation depends on the channel disparity between paired users, which fixed-antenna deployments can only exploit as they find it. Pinching antenna systems turn that disparity into something the network designs. We formulate the joint selection of user pairing, antenna assignment, and intra-group power allocation, and show the power split separates across groups while pairing and assignment do not, because the assignment itself decides the weak-strong ordering. The resulting Q-learning agent engineers higher disparity than proximity, random, and fixed-antenna assignment at equal radiated power, reaching 98.9% of the utility of exhaustive search.
Innovations in Wireless Intelligent 6G Networks
Research Poster, Santa Clara University
Poster on the antenna and multiple access technologies 6G needs to reach 10 million devices per square kilometer, and the barriers still in the way.