PhotonVision
C++ and Java post-processing and quantization for YOLOv8/v11 on CPU, plus a Jupyter conversion toolchain, so FRC teams can run real-time detection on Orange Pi and Raspberry Pi hardware.
- C++
- Java
- YOLO
- CVAT
About
I like building things that have to work in the real world — computer vision on cheap hardware, security holes in production apps, and websites people actually use. I am also curious about how business processes, people, and org systems fit together, and I want to bring a technical lens to that side of the work.
- School
- University of North Carolina at Chapel Hill
- Degrees
- B.S. Computer Science · B.S. Applied Math · Minor in Linguistics
- GPA
- 3.9 / 4.0
- Graduation
- May 2029
Experience
-
Aug 2024 – Present
Open Source Developer / ML Engineer · PhotonVision
- Extended the core computer vision engine with C++ and Java tensor post-processing and quantization for YOLOv8 and v11 on edge hardware.
- Curated datasets and trained custom YOLO models used by 1,000+ FRC robotics teams.
-
Aug 2025 – Present
Autonomy / Perception Engineer · RAM Robotics Club (UNC)
- Built perception for an autonomous delivery robot navigating dynamic environments with ROS 2, Python, C++, and OpenCV.
- Paired Segment Anything sidewalk segmentation with YOLO and LiDAR for obstacle avoidance and spatial awareness.
-
Feb 2026 – Present
Web Developer · Kepha Design
- Turned agency designs into responsive production sites with React, Tailwind, and Astro; shipped 2 client sites on Cloudflare Pages, with a third in progress.
- Wired up contact forms and FullStory analytics to track conversion and surface friction points.
-
Aug 2023 – Aug 2024
SWE Intern (Security + Fullstack) · NUUBI
- Pentested a React, Express.js, and PostgreSQL chat app with 200+ beta users and found broken access control.
- Refactored API authorization to enforce resource access policies and close IDOR holes across chat and forum features.
-
Aug 2023 – Aug 2025
Fullstack Engineer and Lead Robot Programmer · YETI Robotics
- Engineered a Next.js / TypeScript / PostgreSQL scouting platform used by 100+ FRC teams across North Carolina.
- Programmed arm, shooter, and elevator mechanisms in Java with PhotonVision for automated target tracking.
Projects
YETI scouting platform
Next.js, TypeScript, and PostgreSQL robotics scouting app on DigitalOcean, used by 100+ FRC teams in North Carolina to collect and share match data.
- Next.js
- TypeScript
- PostgreSQL
EcoHome Solutions
TSA state-finalist site that helps homeowners find green-energy incentives. Next.js and MongoDB, with Google Maps for geolocation search.
- Next.js
- MongoDB
- Maps API
Taj
Award-winning TSA restaurant site for authentic Indian cuisine, built to show a full marketing page with menu, reservations, and a sustainability story.
- Next.js
- Tailwind
AI UNC wellness predictor
Multi-modal health and wellness forecast combining Fitbit telemetry with patient-reported data. LSTM models in PyTorch plus XGBoost and AHP ranking, with 12% RMSE on an aggregated wellness score.
- PyTorch
- LSTM
- XGBoost
- Pandas
Skills
Languages
- Python
- TypeScript / JavaScript
- Java
- C++
- SQL
- Kotlin
Frameworks & libraries
- Next.js
- React
- Astro
- PyTorch
- TensorFlow
- Django
- Flask
Databases & infra
- PostgreSQL
- MongoDB
- Docker
- Linux
- DigitalOcean
- Cloudflare Pages
Specialized
- Computer vision (YOLO, CVAT)
- ROS 2
- LiDAR
- Security (pentesting)