Fatih Akyon
AI Engineer at Ultralytics
Ankara, Ankara, Turkey
Summary
🤩
Rockstar🎓
Top SchoolFatih Akyon is an AI engineer with 8 years of experience who currently develops production-grade RAG/LLM and agentic AI systems at Ultralytics and co-leads multimodal backbone work on flagship YOLO releases. He co-founded Viddexa and spearheaded video-RAG products and patent filings, securing significant startup support including pre-seed investment and large cloud credits from AWS, Microsoft, and Google. His background spans defense-grade signal processing and academic research (PhD work) through industry delivery—shipping object detection, segmentation, and real-time tracking pipelines that have won competitions and improved F1/AP substantially. An active open-source contributor, Fatih extended Hugging Face’s transformers with a TimeSformer implementation and bootstrapped Sahi’s tiled inference stack, showing a blend of novel model development and practical tooling. He also builds internal tooling that has saved hundreds of weekly human-hours via AI-assisted data labeling and business intelligence automation. Colleagues describe him as a builder who turns cutting-edge research into robust, deployable systems across video, signals, and multimodal AI.
8 years of coding experience
10 years of employment as a software developer
Fen Bilimleri, Fen Bilimleri at Ankara Atatürk Anadolu High SchoolMaster of Science (M.S.) Electrical and Electronics Engineering, Master of Science (M.S.) Electrical and Electronics Engineering at Bilkent UniversityDoctor of Philosophy - PhD Graduate School of Informatics, Doctor of Philosophy - PhD Graduate School of Informatics at Orta Doğu Teknik Üniversitesi / Middle East Technical University
Turkish, English, German
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (14)
multi-object-tracking10
object-tracking10
object-detection10
computer-vision10
transformer10
pytorch10
machine-learning10
yolov510
deeplearning-ai10
deep-learning10
instance-segmentation10
python10
hub9
nlp9
Programming languages (15)
Github contributions (5)
obss/sahi
Jan 2021 - Jan 2023
Framework agnostic sliced/tiled inference + interactive ui + error analysis plots
Role in this project:
Back-end Developer
Back-end DeveloperContributions:108 releases, 321 reviews, 422 commits in 1 year 11 months
Contributions summary:Fatih uploaded the initial code for the project, introducing the core components and functionalities. They implemented an instance segmentation model and added support for basic inference along with the integration of error analysis plots and a UI. The code contributions include the instantiation of a detection model, image loading, and the performance of predictions, laying the groundwork for the framework agnostic slicing/tiling inference.
pythondetectron2tiledagnosticerror-analysis
tryolabs/norfair
May 2021 - Nov 2022
Lightweight Python library for adding real-time multi-object tracking to any detector.
Role in this project:ML Engineer
ML EngineerContributions:9 reviews, 7 commits, 7 PRs in 1 year 6 months
Contributions summary:Fatih primarily focused on integrating and improving a YOLOv5-based object tracking demo within the norfair framework. They implemented a demo for tracking cars and pedestrians, addressed a typo, and corrected an index bug in the YOLOv5 demo notebook. Additionally, the user contributed to the project by adding support for a newer version of Sahi and including versioning information.
pythonre-idreal-timeobject-detectiondetector
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.