About
I'm Ivan Bardak, a solution architect with more than 15 years in IT. I design and sell complex systems for enterprises, and lately most of that work involves making AI agents behave like reliable parts of a real architecture.
In numbers
The short version of fifteen years in enterprise IT.
Today most of my work is architecture for AI agents.
Solution design, estimates and bid narratives for enterprise deals in healthcare distribution, energy, retail and logistics.
The combined budget of the transformation programmes I have worked on as an architect.
I lead pre-sales architecture and agentic AI delivery, and grow engineers into solution architects.
Figures as of October 2026. The programme figure is the combined budget of those programmes, not my own revenue.
What I write about
- Agents and platforms
Getting AI agents into production: MCP tool design, delegated access, reliability, and the agents that now rewrite and migrate code. Read the posts.
- GraphRAG and retrieval
When knowledge graphs beat vector search, what they cost to run, and how to build the business case. Read the posts.
- AI transformation
What controlled studies, surveys and company filings say about the effect of AI, and which frameworks help an organisation adopt it. Read the posts.
How I write
Most writing about AI is either a demo or a vendor pitch. I write about the decisions in between. Every post starts from primary sources, such as papers, specifications and company filings, says how much weight each one deserves, and ends with a decision rule or a checklist you can use.
Expect evidence tables, interactive charts with the data behind them, and my own experience from enterprise projects, marked as such.
I'm based in Ukraine.