ML · COMPUTER VISION · AI

From raw data and sensors to production-grade AI.

I'm Valentin. I build systems for computer vision, machine learning and artificial intelligence integration.

01 /

Who I am

I'm an engineer with 20+ years in .NET, from enterprise systems to ML. I was the CTO of a ~30-engineer team; now I've deliberately gone back to building computer vision and machine learning systems hands-on. I work solo and bring in a proven team when the scope needs it. I experiment, keep my own projects running, and advise people building something ambitious.

02 /

Capabilities

SYS.01

Computer Vision / Edge-AI

Real-time object detection and tracking in video. Runs autonomously on edge devices, no cloud - from sensor to production.

YOLOOpenCVONNXTensorRTNVIDIA JetsonRaspberry Pix86 / ARMC++
SYS.02

Predictive Analytics / ML

I build machine learning pipelines. I find the patterns in your data that predict an outcome - demand, prices, risk - and test them on fresh data.

ML.NETFeature EngineeringLightGBMWalk-ForwardHPOPipeline FrameworkBacktesting
SYS.03

AI Integrations / LLM

I embed AI into your product. Speech recognition and synthesis, images and documents turned into structured data, assistants and search over your own data.

LLM APIsRAGEmbeddingsFunction CallingAgentsSpeech-to-Text.NET
03 /

How I work

Thinking ahead

I dig deep into the problem and think several steps ahead: I spot the pitfalls and the fixes before they become your problem.

End-to-end ownership

I keep every stage under control: from the brief and architecture to development, deployment and support. One owner, no handoffs.

Transparent process

You always know what stage the project is at and what's next, in plain language.

04 /

Tech Stack

Core C# / .NET 10, C++17, Python
Computer Vision / Edge OpenCV, TensorRT, ONNX; Jetson, Raspberry Pi, x86/ARM
ML / Predictive PyTorch, ML.NET, LightGBM
AI / LLM LLM APIs - OpenAI, Anthropic, Ollama, Whisper, RAG
Backend / Infra ASP.NET Core, ABP.IO, REST / gRPC, Docker

Initialize project discussion

Describe what you need in plain words. I'll sort out the technical details after the first conversation.