Dmitry Popov
Verified Expert in Engineering
Computer Vision Developer
Canterbury, United Kingdom
Toptal member since September 24, 2025
With a master's and a post-graduate degree in computer vision, Dmitry has 15+ years of embedded systems experience. He has worked for 10+ years in ML/DL and computer vision and has four years of team leadership expertise. Dmitry has authored seven patents and has worked on various projects, including a UAV vision stack, a utility-pole vibrometer, a roulette video reader, and speech technology.
Portfolio
Experience
- Computer Vision - 20 years
- C - 20 years
- Real-time Operating System (RTOS) - 15 years
- Analytical Geometry - 15 years
- OpenCV - 10 years
- Lean - 5 years
- NVIDIA Jetson - 4 years
- PyTorch - 4 years
Preferred Environment
Ubuntu, Visual Studio Code (VS Code), CLion, PyTorch, Hugging Face Transformers, OpenCV
The most amazing...
...solution I've built is a computer vision detector that revolutionized the land-based casino market, establishing a remote-betting niche for live roulette games.
Work Experience
DL Researcher and Edge-AI Engineer
Insurevision.ai
- Performed vision transformers pre-training and multi-head fine-tuning, class- and domain-re-balancing, hard-negative mining, and self-supervised adaptation.
- Scraped 120,000 hours of taxi video for statistical proof of insurance risk-scoring.
- Automated JetPack 5/6 flashing and tooling for QA benches.
- Piloted a 100-car fleet across three cities, working on an end-to-end edge-ML pipeline.
- Leveraged the Streamlit web-app for visuals, leading to successful proof of concepts (PoCs) with one insurer and two investors.
RTOS Software Engineer
Setel
- Worked on x86/ARM RTOS Multex libraries and hardware drivers.
- Handled a remote betting network stack, video streaming over LAN, 2D graphics engines, a GUI library, and color calibration.
- Applied software development skills based on the exokernel proprietary RTOS Multex.
Head of Computer Vision
MicroAvia
- Worked on optical-flow stabilization at a rate of 60 frames per second, leveraging a $5 IMX219 camera module alongside a Jetson Nano for vision-based landing capabilities.
- Built an UAE-specific ChatGPT-resistant recruiting funnel and closed two specialized hires in three months.
- Delivered pre-sale demos that secured the first defence/industrial pilot.
- Filed four UAV CV patents within visual landing, acoustic localization, and de-jello for video stabilization.
Perception AI Manager
Speech Technology Center
- Scaled an AI team from four to 14 members, working on diarisation and a face/voice biometrics suite for face-to-face speech analytics.
- Built the prototype of a multicam/multi-mic spatial speaker-ID for conference rooms.
- Rolled out a Lean DS and Kanban cadence for a remote team, achieving faster feedback loops and customer-driven data flows. Mentored researchers and engineers.
DL Researcher, Speech
Speech Technology Center
- Conducted deep learning (DL) research for automatic speech recognition (ASR), speech recognition grammar specification (SRGS), and large vocabulary continuous speech (LVCS).
- Performed LM tuning for an office voice-assistant PoC, working on acoustic room simulation for far-field ASR.
- Worked on a voice command recognition prototype at pre-release Intel application-specific integrated circuit (ASIC) with six mics.
Embedded Acoustics ML Engineer
The Pole Expert
- Patented a handheld vibration-based pole-health meter, achieving $2 million in revenue and $30+ million in yearly savings.
- Handled signal processing on a microcontroller unit (Atmega128), achieving a one-week battery life using three AA batteries.
- Tuned the natural frequency model of a cantilever beam using genetic algorithms (GA) and SOLIDWORKS simulations.
Computer Vision Engineer
SET-Production
- Patented a roulette winning number computer vision detector, generating $10 million in revenue.
- Introduced "no-more-bets" and "invalid spin" signals, enabling remote real-money betting.
- Built interpretable casino analytics, allowing GLI certification for remote betting and side-games invention and calculation.
- Reworked CV algorithms for the double-zero roulette to expand to the US market.
Experience
World Bronze Medal in Speech Recognition
https://www.chimechallenge.org/workshops/chime2018/papers/CHiME_2018_paper_medennikov.pdfEducation
Master's Degree in Computer Science
Saint Petersburg Polytechnic University - Saint Petersburg, Russia
Certifications
Lean Data Science Practitioner
Lean DS Community
Professional Scrum Master I
AgiliX Consulting
Skills
Libraries/APIs
Accelerometers, PyTorch, OpenCV, Hugging Face Transformers, FFTW
Tools
CLion, NVIDIA Jetson, Kaldi, SOLIDWORKS
Languages
C, Python, C++, Bash, Perl, Embedded C
Frameworks
GStreamer
Paradigms
Kanban, Scrum
Platforms
Ubuntu, Visual Studio Code (VS Code), Embedded Linux, Google Cloud Platform (GCP)
Storage
Amazon S3 (AWS S3)
Other
Real-time Operating System (RTOS), Computer Vision, Facilitation, AI Product Management, Video Processing, Audio, Projective Geometry, Analytical Geometry, Lean, Hypothesis flow, 3D Scene Reconstruction, NVIDIA TensorRT, DSP, Convolutional Neural Networks (CNNs), Hypothesis Flow, Speech-to-Text (STT), Biometrics, Analytics, Tensilica DSP, Intel GNA, Atmel AVR, Patents, Hardware Drivers, Network Stack, UI Libraries, 2D Graphics, Object Detection, Object Tracking, Image Processing, Computer Science, Scrum Master
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring