Farzin Soleymani

Machine Learning Researcher @Helmholtz Munich

Munich, DE

Implemented and evaluated a multi-modal AI framework, integrating large language models (LLMs) and graph neural networks (GNNs), to predict complex biological interactions. This involved pretraining and fine-tuning LLMs on domain-specific data and integrating diverse data modalities- Engineered a novel topology-aware GNN model, CliquePH, to capture higher-order graph information, achieving up to a 31% performance improvement on benchmark datasets- Built and validated a framework for 3D pose estimation and mapping using point cloud representations and optimal transport techniques. This highlights proficiency in processing and analyzing complex 3D data

Sharif University of Technology

Bachelor of Science - BS, Computer Engineering

Technical University of Munich

Master of Science - MS, Electrical, Electronics and Communications Engineering