I'm a Computer Engineering student at UC San Diego on a combined B.S. and M.S. track. I'm focused on software engineering, especially full-stack development, backend services, and systems programming.
At SDSC, I worked full-stack on an internal research cloud platform, building backend APIs, typed data models, and frontend workflow features used by engineering and research teams. Outside of that, I've built projects across the stack, including multi-tenant web apps, LeetCode review tools, C systems projects, and ML pipelines.
I like working through the full shape of a product: understanding the use case, designing the data model, building the API, connecting the frontend, and making sure the final result is reliable and easy to use.
Certified AWS Solutions Architect · Google AI · Snowflake Data Engineering
to
Built full-stack workflows for SDSC's internal research cloud platform, including production interfaces and backend services across 5+ resource domains (users, quotas, storage, networking, compute).
to
Architected a cross-border e-commerce platform, including backend, schema, and API specs across product sourcing, payments, and logistics.
to
Built robotic arm firmware in C/C++ on Arduino to automate fragile perovskite solar film placement for nanoengineering research.
Previously: President of MSA UCSD (300 members, $130K+ budget), Peer Mentor in the CE department (40+ students), Founder & CEO of AUB Media (digital agency, 150M+ organic views, $250K+ revenue), and Webmaster for SASTA - context on request.
Full-stack interview prep app implementing a 3-stage review workflow (learning → practicing → mastered) with confidence-based scheduling and SM-2 ease-factor updates. Ships with an analytics dashboard (activity heatmaps, weekly trends, retention metrics), authentication, per-user tracking, daily goals, and streak logic. Grew to 10 active users organically with zero spend.
Multi-tenant platform with slug-scoped routing and granular RBAC across 6 permission domains (calendar, tasks, resources, finance, and more). Includes an event request approval workflow with budget metadata and calendar linkage, a folder-scoped resource hub, audit logging, and platform-level admin tooling. Session-based auth with OIDC, transactional email via Resend, responsive TanStack Query frontend with Radix UI.
Unix-style shell in C supporting command parsing, piping, I/O redirection, background job control, and signal handling via low-level system calls. Supports both interactive and batch modes.
Multi-threaded chat server in C with concurrent clients, session auth, private messaging, and broadcast. Also exposes an optional HTTP interface for web-based access via socket programming and mutex synchronization.
Custom heap allocator in C supporting first-fit and best-fit strategies, plus introspection tools to detect leaks and fragmentation across allocation patterns.
ML pipeline predicting total workout duration from the first 5 minutes of heart-rate and speed sensor data. Engineered statistical features from sequential time-series and benchmarked Linear Regression vs. Random Forest with MAE / RMSE / R². Random Forest beat the global mean baseline across all sport categories.
IoT outfit recommendation system integrating ESP32, MQTT, FastAPI, and Docker across distributed edge and cloud layers, processing real-time environmental sensor data end-to-end. On-device inference preprocesses data at the edge to cut round-trip latency.
CNN reaching 98% test accuracy on MNIST by optimizing the full training pipeline, including architecture selection, data loading, and hyperparameter tuning. Cut training time 40% via preprocessing and batch optimization without losing accuracy.
Deployed full-stack community platform on MERN with OpenAI's DALL·E API. Users can generate, browse, and share AI-created images in a social feed.