Pavel Sirenko
I am an experienced AI, ML, and computer vision engineer specializing in shipping low-latency computer-vision systems at scale for domains like ADAS, medical, biometrics, and UAV.
About
I am an experienced AI, ML, and computer vision engineer specializing in shipping low-latency computer-vision systems at scale for domains like ADAS, medical, biometrics, and UAV. I am passionate about optimizing edge inference and building large data pipelines to enable continuous retraining and improvements. Currently, I build hybrid Vision-LLM and GenAI pipelines to convert imagery and LiDAR data into photorealistic 3D city models.
I have hands-on experience with NVIDIA Jetson devices, cloud systems, and a wide range of AI/ML frameworks, and have delivered systems with multi-million data points and sub-30 ms inference times. My background also includes biometric backends, UAV monitoring, and medical imaging innovations, supported by strong programming and system design skills.
- 6 years' experience
- Joined August 2025
Industry experience
- Healthcare
- Life Sciences
- Software & Internet
- Telecommunications
- Transportation & Logistics
Work experience· 6 roles
Architected full-stack ADAS platform with OSM integration running GPU inference on NVIDIA Jetson Orin, edge, mobile, and cloud. Built a system harvesting 200k+ images/month (31M+ total) for self-driving applications. Deployed <30 ms OCR to expand U.S. road-sign coverage 7x, integrated Stable Diffusion for license plate and person augmentation improving recall by 8 percentage points, and developed LLM-powered NER service achieving 90% F1 on address and routing parsing.
Built hybrid Vision-LLM/GenAI/Gaussian Splatting pipeline converting street-level imagery and LiDAR into photorealistic 3D city models. Designed 3D point-cloud modality for VLMs improving object-instance extraction accuracy by over 100% using weak supervision. Led a cross-functional squad of 3 people.
Built biometric ML-powered backend supporting over 2 million users with 99.99% uptime, including liveness detection, ID matching, OCR and glare removal processing around 400k documents per month.
Education
Hire Pavel Sirenko today
Post a job and invite Pavel, or contact them directly. No hidden commission.