Dimitri Voytan

Background

About Dimitri Voytan

AI engineer/scientist with 6+ years of experience bridging autonomous agents and scientific machine learning: Current focus – LangGraph agents in production Building LLM-driven agents for industrial workflows—from legal document triage to asset-performance monitoring. I focused primarily on the observability and performance layer (Langfuse) and wire everything together in Python with robust CI/CD. Research & modeling – diffusion, neural operators, PINNs Designed diffusion generative models for inverse problems, with extensions to seismic imaging and other physics-heavy domains. Published work on Neural Operators and Physics-Informed ML to speed up PDE solvers. Roots in high-performance computing Started work in computational physics writing numerical codes in fortran, C/C++ with MPI, OpenMP, OpenACC, and CUDA—now just as comfortable now orchestrating GPU workloads through PyTorch, Docker, and CI/CD pipelines.