I build and run GPU infrastructure for training and high-throughput AI inference.
Head of engineering for a GPU cloud and inference platform. In parallel, deputy head of a laboratory building an internal developer platform and an ML platform. Eight years in infrastructure, the last four in MLOps.
Where I have worked and what I have worked with is on the CV. The longer version is in about.
// writing
Nothing published yet. The first post will show up here.
// research
2026.03
Good-Enough LLM Obfuscation
With Anatoly Belikov. Keeping prompts private on shared accelerators without paying the cost of MPC or FHE. arXiv:2603.05035