San Diego, California, United States
IF YOU PLAY LINKEDIN GAMES, SEND A CONNECTION! >>> Hey! I'm Herman, a Mathematics & Computer Science major at UC San Diego. Check out my portfolio for more insight on my experience and projects! www.hermanhundsberger.com/ In terms of my personal inspiration and motivation for software, gaming has taught me how I approach engineering. I love being immersed within a game and understanding the complexities that exist to get systems in these games running (shoutout TFT), and how they constantly evolve through iteration and refinement. interacting with these systems and watching them become increasingly versatile has sparked in me an endless hunger to grow and to learn new tools and software practices. Feel free to reach out! :) Email: [email protected] Portfolio: www.hermanhundsberger.com/
Summer 2026
Selected among 100+ teams for Roblox's official game development program, receiving exclusive mentorship from Roblox engineers with access to new AI tooling including procedural model generation and platform-native development resources.
• Lead a 15-engineer full-stack subteam developing the rover mission control system for Yonder Dynamics, nationally competing at the University Rover Challenge • Architected major frontend overhaul: migrated build system to Vite, replaced Janus with low-latency WHEP WebRTC streaming, integrated Three.js URDF arm visualization, and refactored Redux toward localized state management • Designed ROS middleware with pilot/spectator publish-handshake to enable safe real-time multi-user rover control • Established documentation standards and structured code review processes across all software subteams • Authored 20+ page onboarding workshop series reducing new developer ramp-up time by 33%
•Engineered a vector-based Leaflet mapping solution for autonomous path tracing that enhanced navigation planning efficiency and enabled real-time rover tracking capabilities, reducing mission planning time. • Redesigned rover live control interface using React, JavaScript, and ROS, improving UI clarity and reducing operator cognitive load during remote missions • Built system version tracking pipeline that scans all rover repositories at boot, publishes commit metadata via ROS, and surfaces live mission state in Redux
• Engineered a high-availability NFC check-in platform for 8,000+ UCSD students, eliminating manual verification bottlenecks and achieving 99.9% uptime via distributed design and automated failover. • Refactored legacy codebase into modular, thread-safe components and redesigned check-in logic to prioritize local-first execution, further optimized for concurrency across Raspberry Pi nodes. • Improved average check-in speed by 92% through local caching, asynchronous queuing, and deferred synchronization to Google Sheets, enabling real-time LED confirmation during high-traffic periods. • Built a queue-based background thread system with write delay invariants, syncing local check-ins to the online database only after inactivity to prevent race-conditions and duplicate writes. • Implemented secure export pipeline with timeout-protected API calls and verbose logging to ensure fault tolerance.
• Designed and implemented a distributed, IoT-driven behavioral research platform integrating multi-sensor inputs, real-time event processing, and automated hardware control to enable scalable neuroscience experiments on reward-driven learning and social interaction. • Enabled high-throughput, reproducible behavioral studies via fault-tolerant edge data pipelines and automated closed-loop control, improving data integrity across distributed chambers. • Developed a fault-tolerant data architecture with local SQLite buffering, automated rsync synchronization, and JSON event streaming, achieving 100% data fidelity across distributed experimental chambers. • Integrated automated hardware control systems managing pellet dispensing, sensor calibration, and chamber operations through Python-based event handling, reducing manual intervention by 75% during experiments.