Arthur Garzon
The world isn’t going
to build itself.
Scroll↓
About
Here to build what keeps America ahead.
Deep technical background in unmanned systems, computer vision, and autonomous platforms. Driven to build the things that matter — at the intersection of hard engineering and real-world impact.
Experience
Work & Projects
Internships
AI/ML Engineering Intern
Zone 5 Technologies
San Luis Obispo, CA
- —Led integration and debugging of sensor payload for Paladin, a Group 2 UAS, enabling C-UAS flight capabilities in GPS-denied environments.
- —Developed Python scripts to deduplicate YOLO datasets by 10% using fastdup, and converted bounding boxes to polygons with SAM2 for improved segmentation training.
- —Fine-tuned YOLOv8 on 60,000+ UAS images, improving target localization by 9.5% and setting the standard for Paladin's vision system.
- —Updated vision codebase for YOLO segmentation on NVIDIA Jetson; managed GPS-denied operations via SSH/CLI at Camp Roberts.
YOLOv8SAM2PythonNVIDIA JetsonC-UASGPS-denied
Projects
Team Lead — Fire-mapping VTOL Drone
UCSB Engineering Capstone
- —Won the CNSI Innovation grant and $3,000 in funding to create and lead an engineering capstone team.
VTOLFire-mappingTeam LeadCNSI Grant
Counter-UAS Payload
Independent Project
- —Developing a modular payload for PX4/ArduPilot FPV drones enabling kinetic interception for C-UAS operations.
- —Prototyped UAS detection with a fine-tuned YOLOv11n model achieving mAP50 >80% on Raspberry Pi and Jetson Orin Nano.
- —Validated market needs with a U.S. Army Colonel and the sUAS Division Chief, ACM-UAS.
YOLOv11PX4ArduPilotRaspberry PiJetson OrinC-UAS
Real-Time Drone Meeting Platform
Independent Collaboration
Santa Barbara, CA
- —Built a WebRTC web app with livekit-server-sdk for low-latency video streaming from drone platforms.
- —Integrated with PencilSpaces for virtual meeting participation via drone video feeds.
- —Pitched prototype to a venture capitalist, receiving positive feedback on remote surveillance potential.
WebRTCLiveKitFlaskRaspberry Pingrok
Home Surveillance Drone
Independent Project
- —Designed and 3D-printed two indoor drone prototypes with FEA validation in SolidWorks.
- —Developed pymavlink scripts for Raspberry Pi ↔ PX4 integration enabling autonomous indoor patrols.
- —Built a Flask GCS web app for drone monitoring and control.
pymavlinkPX4FlaskRaspberry PiSolidWorks3D Printing