Veli Ates

I’m a machine learning engineer completing my M.Sc. in Embedded Systems Engineering, with a strong focus on computer vision, responsible/fairness-aware AI, and data-centric ML systems. I’ve built tools and workflows that help teams audit datasets, validate assumptions, and make evidence-based training decisions.

In my Master’s thesis work at DFKI, I authored DiversityLens, a PyPI-published Python library for analyzing demographic diversity in robot-human interaction datasets. I processed ~310k images across 10+ public datasets using OpenCV, DeepFace, and RetinaFace, and I designed extensible, pluggable dataset-auditing pipelines with validation checks and structured exports to support reproducible fairness analysis.

I also bring practical data-curation experience from content moderation, where I reviewed classification edge cases for 95%+ policy accuracy—working through labeling ambiguity, annotation taxonomies, and high-volume quality review. Alongside this, I develop applied AI prototypes such as a privacy-preserving local RAG system using Llama 3-ChatQA, FAISS, and SentenceTransformers, and I care deeply about testing, CI/CD, and maintainable engineering.