Building digital employees that deliver outcomes when plans change.
I'm a co-founder of 8Hats Lab. I build digital employees that turn company knowledge into work a team can delegate—with explicit authority and checks for performance, reliability, and alignment.
I work with agents every day, designing how people and agents share work, judgment, and responsibility.
- Open Digital Employee — I maintain this experimental specification for describing digital employees: measurable outcomes, expertise, operating context, permissions, and professional judgment. Includes schemas, synthetic examples, and structural checks.
At 8Hats Lab, we run experiments across agent learning and alignment, shared work between people and agents, and how organizations turn experience into reusable knowledge. We build tools, test assumptions, and develop methods for putting agents to work in real organizations.
Public examples include our Agents Coworking plugins and BoK Fidelity Gate, a reference implementation for reviewing evidence before admitting claims into an agent's knowledge base.
Explore more of our research, products and experiments, and public repositories on the lab's website and GitHub.
- Human context: What does an agent believe about its user, where did that belief come from, and has the person accepted it?
- Agent evaluation: Are we measuring structural compliance, answer quality, or fit between the agent's role and the task?
- Decisions under uncertainty: When should an agent act, ask, wait, or stop as facts and priorities change?
- Computer science — My foundation in software and systems: making ideas executable, understanding constraints, and testing behavior.
- EdTech — Designing how people learn and develop capabilities. I founded MIPT's Online Education Center and launched Russia's National Open Education Platform. My team at the E-learning Development Fund produced 150+ courses for Coursera.
- Transpersonal psychology — My study of agency, intention, meaning, and relationships, alongside Buddhist and shamanic traditions. It informs the questions I ask about how people relate to agents and which choices should remain with the person.
These three strands come together in my work with agents: building the systems, designing how they learn and develop capabilities, and understanding the people they work with.
I want AI to expand human capability and agency: our ability to understand, choose, and act.