Jake Weber

Computer Vision M.S. student at UCF · medical imaging, 3D vision, robotics

Orlando, FL

GitHubLinkedInEmail

I'm a Computer Vision M.S. student at the University of Central Florida (B.S. Computer Science, 2026). I like building vision systems end to end: fine-tuning and evaluating models on GPU clusters, then shipping them as real, containerized applications people can use.

My work spans medical imaging (nuclei segmentation, virtual staining, multimodal question answering on scans), 3D and geospatial vision (cross-view geo-localization, Gaussian splatting), and robotics (a lunar-rover motor drive for the Florida Space Institute and an autonomous quadcopter stack). I care about the engineering around models as much as the models themselves: reproducible experiments, efficient fine-tuning, and infrastructure that doesn't fall over.

Projects

Med-VQA

Live

Oct 2026

Upload an X-ray, CT, MRI, DICOM or histology image, ask a question in plain English, and get a conversational answer grounded in NIH sources with numbered citations and similar real past cases.

75% VQA-RAD yes/no accuracy (live pipeline)

FSI Lunar Rover Motor Drive

Aug 2025 – Apr 2026

Replaced the stepper-motor drivetrain of RE-RASSOR, a variant of NASA's RASSOR lunar excavation rover, with a brushless (BLDC) system: about 81% lighter, about 82% lower power per motor, and more torque.

~81% lighter drivetrain

WRIVA CVGL Challenge

May 2026 – Aug 2026

Located ground-level photos on Maxar satellite imagery (latitude, longitude, heading, pitch) with a customized DINOv3 pipeline, finishing 15th globally in IARPA's 2026 challenge.

15th global leaderboard rank

~40%lower FID vs baseline

Diffusion Virtual Staining

Mar 2026 – May 2026

Compared four diffusion architectures, including three with novel components, for computationally turning routine H&E histology slides into IHC stains on the BCI breast-cancer benchmark, trained on NVIDIA A100 GPUs.

LoRA Nuclei Segmentation

Apr 2026

Adapted the lightweight MobileSAM segmentation model to find cell nuclei in H&E histology by training only 29.7k LoRA parameters (~0.3% of the model), with no prompts needed at inference. It now powers Med-VQA's nuclei-analysis tool.

0.72 Dice

6flight states

Autonomous Drone

In progress

May 2026 – Present

A simulated quadcopter that flies a full mission (arm, takeoff, hover, cruise, land) and holds position when its YOLOv8 vision pipeline spots a person or obstacle. Built as a testbed for flight code headed to real hardware.

Education

  • M.S. Computer Vision

    University of Central Florida

    Expected May 2028 · GPA 3.572

  • B.S. Computer Science

    University of Central Florida

    May 2026 · GPA 3.572

Relevant coursework

  • Medical Image Computing
  • Robot Vision
  • Machine Learning
  • Computer Vision Systems
  • Introduction to Robotics
  • Computer Architecture
  • Operating Systems
  • Systems Software
  • Parallel Computing
  • Linear Algebra

Skills

Languages
  • C
  • C++
  • Python
  • Java
  • JavaScript
  • HTML/CSS
  • Dart
  • LaTeX
Machine Learning & AI
  • PyTorch
  • TensorFlow
  • CNNs
  • Model Compression (LoRA)
  • Deep Generative Models
  • Distributed Inference
  • Vision-Language Models
  • RAG
Computer Vision & Robotics
  • OpenCV
  • MobileSAM
  • YOLO
  • DINOv3
  • ROS 2
  • Gazebo
AI Infrastructure & Tools
  • CUDA
  • Slurm HPC
  • Linux
  • Docker
  • Git
  • Arduino
  • Raspberry Pi

Certifications

  • Certified Kubernetes Application Developer (CKAD) · Expected Feb 2027
  • AWS Certified Machine Learning Engineer – Associate (MLA-C02) · Expected May 2027

Leadership

  • President, UCF Table Tennis · Aug 2024 – Apr 2026

    Led a 387-member organization; managed a $9,500 annual budget for equipment, travel and regional events.