The
Architect
I design AI systems that think, coordinate, and act — then I put them into production.
That's the short version. The longer one: I'm an AI Solutions Architect and consultant who has spent the last five years at the edge where research meets deployment.
I work with enterprise clients across banking, insurance, telecom, and government to build the kind of intelligent systems that don't just answer questions — they reason across tools, orchestrate other agents, and make decisions in real time.
I grew up in Lviv, studied in Paris, and now work from Heidelberg, but the real throughline of my career is one obsession: making AI that actually works under pressure, at scale, in the messy real world.
My core work today sits at the frontier of agentic AI — systems where multiple specialized agents collaborate, delegate, and self-correct to solve problems no single model can handle alone.
I architect multi-agent orchestrations for enterprise: pipelines where a retrieval agent pulls from graph-based knowledge bases, a reasoning agent evaluates options, and an action agent executes — all governed by safety rails and human-in-the-loop checkpoints.
I've shipped a conversational AI system that handled 1.5 million requests in a single day. I've built an LLM-powered engine that generates localized content across 20 languages for a global financial institution. I've designed IoT-to-intelligence pipelines that estimate ship emissions from raw sensor data.
These aren't prototypes sitting in notebooks — they're production systems making real decisions for real organizations.
The
Systems
Builder
The
Visual
Intelligence
Researcher
Before I became an architect, I was deep in visual intelligence — and that foundation shapes everything I build.
At the Paris Brain Institute, I developed a novel GAN architecture for normalizing brain histology scans used in Alzheimer's research. At CEA-List, France's atomic energy commission, I worked on trusted computer vision and robust object detection under conditions I still can't discuss.
I've built depth estimation systems for thermal cameras, fire detection through sensor fusion of RGB and thermal streams, and real-time computer vision for autonomous drones — work that won a grand prize in Abu Dhabi.
I've published on self-supervised learning, political speech emotion recognition, and open-set object detection at venues like IEEE Access, LREC, and ECCV.
Today, I bring that visual intelligence expertise into multimodal AI systems — agents that don't just read text but see, interpret, and reason over images, documents, and video.
I don't just build — I advise. A significant part of what I do is helping organizations figure out what AI can actually do for them.
I cut through the noise of hype to find the architecture that fits. I've led transformation engagements where legacy systems became AI-powered products. I've audited existing AI stacks and rebuilt them for reliability and scale.
I run workshops that turn skeptical engineering teams into confident AI practitioners — and those workshops have repeatedly turned into six-figure project contracts.
When I sit across from a CTO or a VP of Engineering, I speak both languages: I can whiteboard a multi-agent orchestration pattern, and I can explain why it matters to the business in the same sentence. That bridge between deep technical craft and strategic clarity is where I live.
The
Consultant
The
Human
I speak four languages — Ukrainian, English, French, and German.
I hold a Labex scholarship from France and a master's from Paris-Saclay where I graduated top 3. My thesis ended up on the Google Developers blog and became one of the most-viewed posts on their official Instagram.
I co-founded two startups before I was 22, which taught me more about negotiation, ambiguity, and building from zero than any degree could.
And when I'm not designing agent architectures or reviewing pull requests, I cook — seriously.
There's something about the precision of a recipe and the instinct of improvisation that mirrors how I think about AI: you need the science, but the best systems are built by people who know when to trust their judgment over the textbook.