Things I've built

Distributed key-value store with replication across multiple servers, crash recovery, and concurrent client handling. Detailed design decisions and optimizations in documentation.

Multi-stage database with in-memory structures and persistent storage. Buffer pools with LRU eviction, static B-Trees, and an LSM-tree with Bloom filters to reduce disk I/O.

A parallel memory allocator focused on speed, scalability, false-sharing avoidance, and minimal fragmentation. Includes design and performance evaluations.

User-oriented system that helps clinicians design rehabilitation curriculums tailored to individual patients — placing virtual objects in the patient's living areas to define tasks without code.

A writing assistant that suggests improvements while preserving the writer's voice — instead of fully rewriting content with overly fancy AI-generated prose.

Web app with user registration, admin governance, profile management, interactive forums, and game performance tracking with stats on win rates and champion usage.

Investigated MINGAR's Fitness Tracker product line and potential wearable device deficiencies.

Examined factors influencing variation in admission rates across U.S. colleges and universities.

Linear regression model relating MLS player earnings to performance — useful for clubs estimating value and players targeting earnings growth.

Identified specific tasks that drive high CPU usage to help prevent slowdowns and crashes.

Designed six utility functions and applied them to 80,000+ users and 12.7M tweets, expanding online communities through new algorithms. Presented at the 2022 ROP Fair.