Open source

My open-source computer vision work, in the open

Most of what I build ships in public. My YOLO detection, tracking, segmentation and pose-estimation projects have earned 2,176 GitHub stars from engineers who put them to work. On top of that I’ve landed 735 merged pull requests across open source, including 464 into Ultralytics YOLO— so the tools I recommend are ones I actually help build.

View my GitHubWork with me
2,176
Stars earned
Across the 45 open-source projects I've built and maintain.
735
Merged pull requests
Shipped across open-source computer vision.
464
Merged into Ultralytics YOLO
Features, fixes and docs on the core repo.
1,100+
GitHub followers
Engineers who follow the work as it ships.

Figures from the GitHub API, snapshot July 2026.

Star growth across my repositories

Cumulative GitHub stars on the projects I own, reconstructed from every stargazer’s timestamp. The jump in 2022 is when my real-time YOLOv7 tracking, pose and segmentation repos took off, and the curve has climbed steadily ever since.

2,176 stars and counting
5001,0001,5002,0002,5002020202120222023202420252026
Cumulative GitHub stars across my repositories by month
Oct 20201
Nov 20203
Dec 20203
Jan 20213
Feb 20213
Mar 20213
Apr 20213
May 20214
Jun 20215
Jul 20215
Aug 20215
Sep 20215
Oct 20215
Nov 20215
Dec 20215
Jan 20225
Feb 20225
Mar 20227
Apr 20229
May 20229
Jun 20229
Jul 20229
Aug 2022218
Sep 2022344
Oct 2022419
Nov 2022526
Dec 2022654
Jan 2023801
Feb 2023885
Mar 2023982
Apr 20231,043
May 20231,097
Jun 20231,142
Jul 20231,200
Aug 20231,253
Sep 20231,300
Oct 20231,335
Nov 20231,378
Dec 20231,410
Jan 20241,473
Feb 20241,497
Mar 20241,536
Apr 20241,566
May 20241,585
Jun 20241,606
Jul 20241,637
Aug 20241,652
Sep 20241,677
Oct 20241,733
Nov 20241,774
Dec 20241,799
Jan 20251,826
Feb 20251,846
Mar 20251,865
Apr 20251,874
May 20251,907
Jun 20251,915
Jul 20251,948
Aug 20251,961
Sep 20251,986
Oct 20251,999
Nov 20252,016
Dec 20252,062
Jan 20262,090
Feb 20262,102
Mar 20262,127
Apr 20262,149
May 20262,161
Jun 20262,169
Jul 20262,176

My work on Ultralytics YOLO

My biggest single body of open-source work is Ultralytics YOLO, where I’ve landed 464 merged pull requests— working across the detection, tracking, segmentation and solutions code, the documentation, and the example notebooks people use to get started.

That means when you read the YOLO docs, run a solutions example, or use an object counting, speed estimation or heatmap feature, there’s a good chance you’re touching something I helped build or write. It’s also why my advice on detection, tracking and deployment comes from the source, not from the outside.

See my merged PRs
In the last 12 months
756
Contributions
225
Commits
219
Pull requests
16
Code reviews

Popular repositories

Open-source projects I maintain, from real-time object tracking to instance segmentation, pose estimation and the latest segmentation models. Each one ships runnable code.

yolov7-object-tracking

Real-time object detection and multi-object tracking with YOLOv7, OpenCV and SORT. One of the most-forked YOLOv7 tracking repos on GitHub.

653172Python

yolov7-pose-estimation

Human pose estimation and keypoint detection on live video using YOLOv7, OpenCV and PyTorch.

37988Python

yolov8-object-tracking

Detection, tracking and counting with Ultralytics YOLOv8, OpenCV and PyTorch, ready to run on your own footage.

37166Python

yolov7-segmentation

Instance segmentation with YOLOv7, drawing per-object masks on images and video with OpenCV and PyTorch.

34381Python

yolov5-object-tracking

YOLOv5 detection, tracking and object blurring with a Streamlit dashboard built on OpenCV and PyTorch.

12428Python

sam3-inference

Run inference with Meta's Segment Anything Model 3 (SAM 3) for promptable image and video segmentation.

6912Python

yolov7-object-blurring

Automatically blur detected objects, like faces or plates, in images and video using YOLOv7 and OpenCV.

6518Python

trajectory-forcast

A lightweight Ultralytics YOLO extension that adds Kalman-filtered trajectory forecasting to predict where tracked objects move next.

197Python

streamgrid

Multi-stream video inference with Ultralytics YOLO, displaying many camera feeds in a real-time detection grid.

163Python
Browse all repositories

Also contributed to

Beyond my own repos, I’ve had pull requests merged into other widely used computer vision and edge-AI projects.

ultralytics/ultralyticsultralytics/yolov5ultralytics/yolov3ultralytics/notebooksultralytics/assetsmemryx/MemryX_eXamples

Want the person behind the pull requests on your project?

I help teams ship computer vision to production, from a first prototype to a system running on cloud, on-prem or the edge. Book a free call and tell me what you’re building.

Book a consultationRead the blogSee the projects