Sartaaj Khan

Chemical Engineering & Applied Chemistry · University of Toronto

Advancing data-driven materials for climate and energy.

I am a PhD candidate in the Department of Chemical Engineering and Applied Chemistry at the University of Toronto. My work involves building geometry-aware representations of metal-organic frameworks (MOFs) in machine/deep learning. If you ever want to talk about AI, chemistry, material science, computer science or just about life, feel free to contact me!

Google Scholar GitHub LinkedIn CV BlueSky

🧪 Research Interests

Multimodal ML figure
2025 • Connecting metal-organic framework synthesis to applications using multimodal machine learning

Screening for MOF applications in sustainable energy

We want to leverage data readily available upon MOF synthesis and use that to directly recommend MOFs for applications such as carbon capture, gas storage and semiconductors.

Multimodal ML Water stability Carbon capture XRD
Topology/structure graphic
2025 • Capturing non-local features of crystals from their bond networks

Geometry-aware representations for porous crystals

A big challenge in existing methods is capturing long-range interactions (and geometry) of porous crystals. Building new representations and algorithms that can capture this aids in the prediction of geometry-reliant MOF properties such as mechanical properties, gas separation and charge transport.

DescriptorsGNNTopology
MOF database pipeline diagram
2025 • MOF-ChemUnity: Unifying MOF data using large language models

MOF Database Development

Using large language models (LLMs), it is now possible to use extraction methods to construct MOF databases that link experimental and existing computational methods together. However, proper curation of these databases is needed for the development of benchmarks to build robust ML models.

Large language model (LLM)BenchmarksCuration

🚀 Publications

Please refer to my Google Scholar for my full list of authored works.

  1. & (2025). Connecting metal-organic framework synthesis to applications using multimodal machine learning. Nature Communications, 16, 5642.

    Open access Code

    Khan, S. T., & Moosavi, S. M. (2025). Connecting metal-organic framework synthesis to applications using multimodal machine learning. *Nature Communications, 16*, 5642. https://doi.org/10.1038/s41467-025-60796-0

  2. & (2025, March). Capturing global features of crystals from their bond networks. AI for Accelerated Materials Design (ICLR 2025).

    Co-first author (equal contribution).
    Open access Code

    Ai, Q., Khan, S. T., Barthel, S., & Moosavi, S. M. (2025, March). Capturing global features of crystals from their bond networks. *AI for Accelerated Materials Design (ICLR 2025)*. https://openreview.net/forum?id=wLSmBbYDY5

  3. & (2025). MOF-ChemUnity: Unifying metal-organic framework data using large language models. Under review in Journal of American Chemical Society.

    Open access Code

    Pruyn, T. M., Aswad, A., Khan, S. T., Black, R., & Moosavi, S. M. (2025). MOF-ChemUnity: Unifying metal-organic framework data using large language models. *Preprint*.

  4. …, …, & (2025). 32 examples of LLM applications in materials science and chemistry: Towards automation, assistants, agents, and accelerated scientific discovery. Machine Learning: Science and Technology.

    Open access Code

    Zimmermann, Y., Bazgir, A., …, Khan, S. T., …, & Blaiszik, B. (2025). 32 examples of LLM applications in materials science and chemistry: Towards automation, assistants, agents, and accelerated scientific discovery. *Machine Learning: Science and Technology*.

  5. … & (2025). Thermodynamics-informed machine learning for predicting temperature-dependent chemical properties. Preprint.

    Open access Code

    Kochi, M. R., Rezaei, H., Khan, S. T., Mamillapalli, B. T., Ebrahimiazar, M., Ye, H., … & Moosavi, S. M. (2025). Thermodynamics-informed machine learning for predicting temperature-dependent chemical properties. *Preprint*.

  6. …, … & (2024). Reflections from the 2024 large language model (LLM) hackathon for applications in materials science and chemistry. arXiv preprint (arXiv:2411.15221).

    Open access

    Zimmermann, Y., Bazgir, A., …, Khan, S. T., … & Blaiszik, B. (2024). Reflections from the 2024 large language model (LLM) hackathon for applications in materials science and chemistry. *arXiv* preprint (arXiv:2411.15221). https://arxiv.org/abs/2411.15221

🏆 Awards

Selected scholarships, awards and honours.

🤖 Software and Datasets

Open-source tools and datasets supporting the community.

MOF recommendation system

XRayPro

A recommendation system for MOFs leveraging only PXRDs and precursors.

Physics-informed ML

ThermoML

Thermodynamic-informed model for predicting thermophysical properties of fluids.

MOF database

MOF-ChemUnity

A knowledge graph unifying computational and experimental data for MOFs.

Solution chemistry

pySolution

A solution chemistry toolkit that computes solution characteristics for modeling purposes.

🎓 Teaching

Courses and instructional roles.