Kareem Abdelmaqsoud

Kareem Abdelmaqsoud

PhD Student at CMU

About Me

I am a 5th-year PhD student in Chemical Engineering at Carnegie Mellon University, co-advised by Professors John Kitchin and Andrew Gellman. My research uses AI to accelerate materials and chemical discovery, combining Density Functional Theory (DFT), machine learning interatomic potentials, and LLM agents. I have applied these tools to catalysis, the design of ductile refractory alloys validated with experimental collaborators, and autonomous optimization of electrochemical interfaces. I also collaborate with Meta’s Fundamental AI Research Chemistry (FAIR-Chem) team on large-scale universal models for atoms.

Industry Experience

Bosch USA — Computational Materials Science Intern (May – August 2026, Watertown, MA)

  • Applied universal ML potentials to build a phase diagram of the Aluminum–Germanium (AlGe) alloy for semiconductor applications in MEMS devices.
  • Studied the effect of Si contamination on the phase stability and eutectic composition of the AlGe alloy.

Orbital Materials — Machine Learning Researcher Intern (May – August 2025, Princeton, NJ)

  • Developed OrbMol, a large-scale ML potential trained on 100M organic molecular structures incorporating magnetism and charge effects.
  • Achieved state-of-the-art accuracy compared with existing molecular ML potentials.
  • Contributed to model architecture, large-scale dataset preparation, and training pipelines.

Entos (now Iambic Therapeutics) — Machine Learning Engineer Intern (January – July 2022, Remote)

  • Performed hyperparameter tuning for the model used by the experimental drug discovery team to improve its accuracy.
  • Implemented new pretraining methods such as contrastive learning that led to a 10–20% improvement in accuracy.

Publications

  • Investigating the Error Imbalance of Large-Scale Machine Learning Potentials in Catalysis (Catalysis Science & Technology) [Royal Society of Chemistry] Abdelmaqsoud, K., Shuaibi, M., Kolluru, A., Cheula, R., & Kitchin, J. R.
  • Structure-Sensitive Reaction Kinetics of Chiral Molecules on Intrinsically Chiral Surfaces (Journal of Physical Chemistry C) [American Chemical Society] Abdelmaqsoud, K., Radetic, M., Fernandez-Caban, C., Widom, M., Kitchin, J. R., & Gellman, A. J.
  • Uncertainty Quantification in Graph Neural Networks with Shallow Ensembles (Machine Learning: Science and Technology) [arXiv] Vinchurkar, T., Abdelmaqsoud, K., & Kitchin, J. R.
  • Computational Design of Ductile Additively Manufactured Tungsten-Based Refractory Alloys (Computational Materials Science) [arXiv] Abdelmaqsoud, K., Sinclair, D., Karra, V. S. S. A., Taheri-Mousavi, S. M., Widom, M., Webler, B. A., & Kitchin, J. R.
  • Electronic Structure and Elasticity of the Ta–W Solid Solution (Physical Review Materials) [arXiv] Abdelmaqsoud, K., Kitchin, J. R., & Widom, M.
  • UMA: A Family of Universal Models for Atoms (NeurIPS) [NeurIPS] Wood, B. M., Dzamba, M., Fu, X., Gao, M., Shuaibi, M., Barroso-Luque, L., Abdelmaqsoud, K., Gharakhanyan, V., Kitchin, J. R., Levine, D. S., Michel, K., Sriram, A., Cohen, T., Das, A., Rizvi, A., Sahoo, S. J., Ulissi, Z. W., & Zitnick, C. L.
  • The Open DAC 2025 Dataset for Sorbent Discovery in Direct Air Capture (Nature Chemical Engineering) [arXiv] Sriram, A., Brabson, L. M., Yu, X., Choi, S., Abdelmaqsoud, K., Moubarak, E., de Haan, P., Löwe, S., Brehmer, J., Kitchin, J. R., Welling, M., Zitnick, C. L., Ulissi, Z. W., Medford, A. J., & Sholl, D. S.

Recent News

  • September 24: I successfully passed my PhD Proposal!
  • August 24: My two papers got accepted!
  • August 23: I successfully passed my PhD Qualification Exam!