I am a Chemical Engineering PhD candidate at LSU in Romagnoli Group utilizing the benefits of physics & machine learning to enhance the design of electrochemical systems.
"Live as if you were to die tomorrow. Learn as if you were to live forever"
- Mahatma Gandhi
- 🔭 Seeking to build data-driven models for real-world applications
- 🌱 Developing Deep Learning and physics-based models for materials (small molecules like solvent, ions; big molecules like surfactant, polymers) and processes (like electrodialysis, electrodeionization, capacitive deionization).
- Chemical Modeling with Physics and Data-driven method
- Chemometrics Data Engineering & Science
- Computational Molecular design
- Material & Process Optimization
- Molecular Simulation of Materials
- Machine Learned Force Field (MLFF) development
- Languages: Python, MATLAB
- Machine learning: Scikit-Learn, TensorFlow, Keras, PyTorch, MLflow, Docker, Streamlit, PySpark
- Chemical Eng. & Chemistry: Aspen Plus, GROMACS, LAMMPS, Gaussian, Rdkit, Deep Graph Library (dgl).
- Platforms: Linux, Git
- Soft Skills: Research, Leadership, Event Management
- Proficiency in the use of Microsoft Office Power Point, Word, Excel, and JMP
- Synthesis & characterization: Nanocrystals synthesis, catalyst synthesis, X-ray diffraction (XRD), Diffuse Reflectance IR Fourier Transform Spectroscopy (DRIFTS), UV etching and Design of Experiment.
- Bridging Physics and Data-Driven: Developed numerical model and data-driven model to perform optimization studies for two common electrochemical systems (electrodialysis and electrodeionization). Under review
- Physicis Informed Machine learning: Developing PINN and Neural ODE to resolve limitations of physics ODEs in capturing selective ion separation in electrodialysis. In preparation
- Transfer Learning for missing data: assess the possibility of resolving missing data with transfer learning. Code
- Feature Embedding: Combined information from experiment, molecular structure and molecular simulation with machine learning to enhance predictive modeling of membrane properties. Code
- Generative Molecular Design: Combined generative artificial intelligence (AI), predictive modeling, reinforcement learning and molecular dynamics to create molecules with desired properties. Code
- Facial Recognition: Collaborated on the development of software utilizing Mediapipe + Blender framework to track facial structure and emotion classification via a trained CNN-based classifier. Code
- Failure detection in pumps: Participated in BPX hackathon and developed a LSTM-based data-driven model to estimate ESP run life. Ranked 3rd out of 30 submissions and received the Implementation award for code reproducibility. Code
- Active Learning modeling: Developed codes to train active learning models based on different query strategies. Presently testing the methods on problems such as protein adsorption, structure-property modelling, & electrochemical separation performance. Code
- Transformer: Trained transformer to encode sequence and classify with PyTorch, & HuggingFace. Code 1 & Code 2
- KNN guided molecular design: Developing a molecular design optimization framework integrating k-Nearest Neighbour and Genetic Algorithms. Code
- Piano Music Generation: Trained two deep learning LSTM models as 1) critic of good or bad music and 2) composer to generate new music. Tools: Python, PyTorch, Scikit-Learn.
- Facial Recognition: Collaborated on the development of software utilizing Mediapipe + Blender framework to track facial structure and emotion classification via a trained CNN-based classifier. Code
- Tox24 challenge: Predict the in vitro activity of compounds from chemical structure. Code
- Website: teslim404.com
- Email: [email protected]
- LinkedIn: Teslim Olayiwola
- Twitter: teslim404
- Google Scholar: Teslim Olayiwola