Products I've built.
Every one shipped from zero. Research to production, product to distribution.
ChaChat
Grew an AI Companion Platform to 3M+ users and $2M+ revenue in 2025.
Luminal.gg
AI workspace for creating and scaling video ad campaigns.
Neiro.ai
Built a generative video platform integrated into Adobe and Canva.
Luratalk
AI tutors for conversational learning — used by thousands of students.
How I work.
I believe speed wins. Ship fast, learn from failure, compound the lessons. I use ML as a lever — not the point — to build products people want.
- Build distributed training and evaluation pipelines end to end.
- Ship AI products used by millions.
I switch between hands-on coding and building teams depending on the biggest bottleneck to the mission, going from a paper on arXiv to a product with real revenue in weeks, not quarters.
Background.
I earned my MSc in Machine Learning at
Skoltech, focusing on Machine Learning algorithms that now power genAI systems. During my studies I also joined the largest private bank in Russia —
Alfabankto build from scratch a transactional credit-scoring system for their very first crowd-lending platform.
After graduating, I joined a selective two-year program at
GE Digitalto ship ML services as cloud products. I left after one year to join
Samsung AI, where I worked on computer vision and generative models.
Two years later, ML models were good enough to become the foundation of consumer AI products. I foundedNeiro.ai, raised capital, and hired a team of over a dozen engineers to build products from that thesis. Some succeeded, some failed — every cycle taught us more about what the market wants.
ChaChat, my most successful product, generated over $2M in 2025 and reached millions of people on their devices. I keep the team lean — fewer than 20 people — and fast-pacing, running 600+ experiments per year to stay focused on the primary goal.