Umut Arisoy

At oNabu, the company I co-founded, I lead product for an AI coaching platform that reads Jira, Azure DevOps, Slack, Teams, and meeting transcripts to deliver feedback where teams work. I shaped its evolution from a rules-based visibility tool into a personalized chain-LLM coaching product used by customers including LC Waikiki, Türkiye İş Bankası, Migros, Bayer, Boyner, and TAV Tech.

I build the evaluation bar for AI products when there is no precedent: defining domain rubrics with subject-matter experts, refining prompts and retrieval through customer feedback, and balancing qualitative insight with adoption, activation, retention, and churn signals in SQL. At LC Waikiki, a deployment across 77 teams and 701 contributors surfaced 3,294 hidden dependencies and increased customer-focus scores by 112% over 110 days.

Before oNabu, I coached 150+ teams and trained 1,500+ people across Scrum, Kanban, Scaled Agile, and DevOps, after a decade building software and delivering large-scale technology programmes. I still work hands-on as a product engineer, using agentic workflows, Claude Code, Codex CLI, Python, SQL, GKE, and Docker to prototype, de-risk, and ship.