Pol Sanz Berman

Computational Chemist & ML Researcher

Tarragona, Spain github.com/pol-sb 0000-0001-8736-404X
Get in Touch

Professional Summary

Predoctoral Researcher in Computational Chemistry specializing in the end-to-end development of tooling and workflows for computational chemistry research, with a focus on machine learning interatomic potentials (MLIP). Expertise in Python-driven scientific computing, data pipelines, orchestration of automated workflows and data visualization. Proficient in managing Linux-based systems and interfacing with HPC environments and modern hardware to scale complex molecular simulations and accelerate catalyst discovery.

Experience

Chemical Data Analyst (Internship)

Cymit Quimica S.L.
  • Designed and implemented a data framework to generate a comprehensive taxonomy of +1M chemical products and their properties, integrating the resulting data infrastructure with the company's ERP, enabling dynamic classification and product categorization on their online marketplace.
  • Programmatically extracted and structured information from heterogeneous sources including chemical databases, OCR parsing tools, web scraping pipelines and on-the-fly molecular mechanics simulations.

Education

PhD in Computational Chemistry

Catalan Institute of Chemical Research (ICIQ) – Universitat Rovira i Virgili (URV)
  • Doctoral Thesis: Metal-oxide and oxide-oxide interfaces for CO₂ recycling. (Supervisor: Prof. Nuria López)
  • Systematically designing machine learning interatomic potentials based on DFT simulations and automated active learning workflows to explore novel catalyst structures for metal-oxide and oxide-oxide interfaces and oxide-derived Cu catalysts for CO₂ reduction.

Master's Degree in Atomistic and Multiscale Computational Modelling

Universitat de Barcelona (UB) – Universitat Politècnica de Catalunya (UPC)
  • Master's Thesis: Modelling small-molecule interactions with intrinsically disordered protein condensates. (Supervisor: Dr. Ramon Crehuet)
  • Studied the influence of small molecules on IDP liquid-liquid phase separation by developing a custom Python-based molecular dynamics framework for coarse-grained protein modeling.

Bachelor's Degree in Chemistry

Universitat de Barcelona (UB)
  • Undergraduate Thesis: Theoretical study of the relative stability of pyrrolidine enamines and their nitro derivatives. (Supervisors: Dra. Anna Maria Costa Arnau and Dr. Jaume Vilarrasa i Llorens)
  • Conducted an in silico study of the position of the equilibrium of several organocatalytic cycles using DFT methods, developing tooling to automate the procedure.

Research & Publications

2026

Li, L., …, Sanz Berman, P., López, N. Dynamic Polaronic Control of Metal Cluster Adaptability on Reducible Oxides. Journal of the American Chemical Society.

2026

Morandi, S., …, Sanz Berman, P., …, López, N. An end-to-end framework for reactivity in heterogeneous catalysis. Nature Chemical Engineering.

2025

Nogueroles-Langa, I., …, Sanz Berman, P., …, Pérez-Ramírez, J. Polyethylene hydrogenolysis to liquid products over bimetallic catalysts with favorable environmental footprint and economics. Nature Communications, 16(1), 9791.

2025

Naeem, M. S., …, Sanz Berman, P., …, López, N. Enhanced proton transport in Fluorine-free membranes by balancing confinement and pore chemistry. ChemRxiv preprint.

2024

Ram, R., …, Sanz Berman, P., …, García De Arquer, F. P. Water-hydroxide trapping in cobalt tungstate for proton exchange membrane water electrolysis. Science, 384(6702), 1373-1380.

Presentations & Posters

Talk

MAI4AM2025 conference in San Sebastian, Spain
Optimizing Active Learning Strategies for Neural Network Potentials in Catalyst Characterization Workflows

    Poster

    MLMMS workshop in Zadar, Croatia.
    Machine Learning Interatomic Potentials for Catalysis

    Poster

    18th ICC conference in Lyon, France
    Active Learning workflows for catalyst characterization

    Poster

    ML4MS Workshop in Ljubljana, Slovenia
    Neural Network Potentials for catalyst characterization

Grants

Predoctoral Grant

PRE2022-101291

Predoctoral fellowship linked to project PID2021-122516OB-I00, funded by the Spanish Ministry of Science and Innovation (MCIN/AEI/10.13039/501100011033) and by the FSE+.

Design

Cover Design

Digital Discovery
Back cover design for Digital Discovery Journal, Issue 6, 2023.

Complementary Education

  • 2026 | Fundamentals of Accelerated Computing with CUDA Python • NVIDIA Deep Learning Institute
  • 2024 | Training on Prevention and Comprehensive Response to Sexual and Gender-Based Harassment • ICIQ
  • 2023 – Present | Severo-Ochoa PhD Training Programme (SHARP) • ICIQ
  • 2022 | CCPBioSim Training Week • Biomolecular simulation workshop.

Technical Skills

Programming Languages

  • Python, Julia, Fortran, C#, Bash

Scientific Codes & Libraries

  • VASP, LAMMPS, AiiDA, OpenMM, MACE
  • PyTorch, NumPy, pandas, Numba, scikit-learn
  • Amber, GROMACS, Gaussian

Infrastructure & Deployment

  • Linux, SGE, Slurm, HPC Administration
  • Bash Scripting, Workflows, Linux SysAdmin
  • Git, Docker, Singularity
  • OpenMP, Open MPI, QEMU
  • Flask, Nginx, Astro

LLMs and Agentic Workflows

  • Ollama, llama.cpp
  • OpenCode, OpenWebUI

Design & Other Tools

  • Blender, Inkscape, GIMP, Photoshop, Illustrator
  • LaTeX, Typist, ChemDraw

Languages

  • Spanish (Native)
  • Catalan (Native)
  • English (C2)*
  • German (A1)

*Official Cambridge English certification, verification number: B7586152.