Dissertation: "Accelerating Theoretical Anharmonic Vibrational Analyses with Machine Learning"
Capstone Thesis: “Evaluation of Deep Eutectic solvents as extraction media for flavonoids from Mahonia aquifolium”
- Developed and implemented predictive machine learning models to accurately predict vibrational properties of molecular crystals 100x faster than quantum-mechanical methods, enabling real-time design feedback.
- Leveraged Python programming skills to automate routine tasks and enhance efficiency of research processes, reducing the time to generate datasets for material property prediction by 75 %.
- Utilized HPC resources to perform complex molecular dynamics simulations and quantum chemistry calculations, significantly accelerating research timelines and enhancing accuracy in predicting chemical behaviors.
- Designed and implemented various hyperparameter optimization strategies to streamline data analysis processes in advanced academic research.
- Presented results/findings at technical conferences and furnished several publications highlighting the application of techniques to simulate material properties.
- Contributed to documentation for the Vermont Advanced Computing Core (VACC), facilitating the use of Intel's oneAPI.
- Utilized distributed training strategies to scale model training across many GPUs on clusters to rapidly train and prototype models
Machine Learning Engineer | Engineers for Ukraine
- Designed and developed ML models to classify logos and insignias in user-uploaded images, increasing the speed at which accurate information about Russian soldiers/equipment in the area passes from local civilians on the ground to Ukrainian soldiers.
- Served as a veteran contributor to PyTorch Lightning with 8 merged PRs deprecating functions/properties and setting up automated tests for the next release version.
- Developed a feature to allow the evaluation of gradients in inference mode, preventing the copying of tensors, resulting in a 10% increase in inference speed.
PyCRYSTAL23 (Demo)
- Python-based webapp deployed on cloud services automating the generation of CRYSTAL23 input files for materials simulations.
- Designed and implemented an intuitive user interface and hosted PostgreSQL database, streamlining the input parameter process and enhancing user experience.
ChemQuery (Github)
- RAG demo running Gemma-7b-it completely locally on an RTX 3070 to answer questions about the CRYSTAL23 manual.
-
"VSCF and VCI Analysis of the Anharmonic Coupling of Stretching Vibrations in Ice XI", Champlain Area Chemistry Symposium, October 2022
-
"LightningANI: A Scalable Template for Building End-to-End Neural Network Potential Applications for Accelerated Material Property Prediction", Northern New England Materials Research Conference, September 2022, Poster Presentation
-
"VSCF and VCI Analysis of the Anharmonic Coupling of Stretching Vibrations in Ice XI", ACS Fall Meeting, August 2022
-
"Utilizing Machine Learning for the Calculation of Vibrational Frequencies of Molecular Crystals", UVM Student Research Conference 2021, Poster Presentation
-
“Anharmonicity of Ice-XI with the VSCF and VCI Models”, UVM Student Research Conference 2020, Poster Presentation
-
Rama Oktavian, Raymond Schireman, Lawson T. Glasby, Guanming Huang, Federica Zanca, David Fairen-Jimenez, Michael T. Ruggiero, and Peyman Z. Moghadam. Computational Characterization of Zr-Oxide MOFs for Adsorption Applications. ACS Applied Materials & Interfaces, 14(51):56938–56947, 2022
-
Raymond Schireman, Jefferson Maul, Alessandro Erba, and Michael T. Ruggiero. Anharmonic Coupling of Stretching Vibrations in Ice: A Periodic VSCF and VCI Description. Journal of Chemical Theory and Computation, 18(7):4428–4437, 2022