Agentic AI

I see Agentic AI as a practical tool for better software work: faster understanding, clearer decision support, and more consistent delivery. It works best when combined with experienced system design and clear quality boundaries.

FOCUS

Agent-based developer workflows

I actively work with how AI agents can support analysis, implementation, testing, documentation, and code review without replacing technical judgment.

MCP and tool integration

Model Context Protocol makes agent workflows practically useful when they are connected to the right tools, data, and constraints. I focus on integrations that create value in real workflows.

A2A and multi-agent systems

I follow Google Agent2Agent and patterns for collaboration between agents, especially where responsibility, traceability, and handoff between roles need to be clear.

LLM security and quality

AI support needs guardrails, testability, and clear boundaries. I work with OWASP perspectives, prompt injection risks, and practical controls for systems that use LLMs.

PRACTICAL USE

  • Faster mapping of codebases and dependencies
  • Better technical decision support and documentation
  • Support for test design, edge cases, and quality assurance
  • Automated workflows around development and delivery
  • Reasonable security boundaries for AI-assisted internal tools

CONNECTION TO MY PROFILE

My background is not AI first and software development second. It is the other way around. I come from backend systems, integrations, cloud platforms, and technical leadership, and I use AI where it genuinely strengthens the work.

That makes me most interested in AI that makes technical teams more effective, not solutions that only look impressive in a demo.

Read about me