Yulian Manchev

I am currently a Computational Scientist at Imperagen, where I develop models to predict enzyme activity, stability, and selectivity. My work sits at the intersection of in silico molecular simulation (including quantum chemistry and molecular dynamics), high-performance computing, and machine learning. I run quantum mechanical calculations and GPU-accelerated molecular dynamics simulations to study biomolecular systems at atomic resolution, and design and process large-scale simulation datasets that feed into machine learning models predicting protein and ligand behavior. I develop computational tools for docking, protein–ligand interaction modeling, and protein structure prediction, contributing to both method development and applied research. Additionally, I design agent-driven workflows to orchestrate complex simulation and analysis pipelines, improving scalability and efficiency across computational experiments.

I hold a PhD in Computational Chemistry, focused on machine learning potentials (MLPs) and next-generation force field development. My research was focused on development of MLPs built on quantum mechanical calculations, and deployment of the models in simulations of biologically relevant molecules. During this time, I was the core developer of ichor Python library which interfaces with high-performance compute clusters and quantum chemistry software, and streamlines the process of dataset creation, dataset processing, and model creation. Previously, I graduated with an MChem degree (First-Class) from the University of Manchester.

I have a deep interest in the intersection of chemistry, physics, and computer science. Particularly, I am interested in machine learning applications in the areas of quantum chemistry, drug discovery and development, protein structure prediction, and materials science.