At Celikkanat AI Technologies, I designed and operate the AI Recommendation Index, an autonomous weekly workflow that processes around 700 LLM responses across eight categories. It uses OpenAI and Gemini APIs to extract and rank brands, track changes, and publish results without manual steps.
I also built FERSAN, a governance layer for Claude Code and Codex that runs agents in isolated, permission-restricted sessions and logs prompts, outputs, and costs. At DowAksa, I worked with operations, quality, and IT experts to replace manual workflows with production ML systems; automation cut data-processing time by 40% and saved more than 20 staff hours per week.
Earlier, at Turkish Land Forces Command HQ, I managed an ML personnel-assessment and workforce-planning system that reduced evaluation time by 50% and increased accuracy by 30%. I bring an M.Sc. in Statistics and experience taking AI and data systems from workflow discovery through production, adoption, and impact measurement.
