I’m drawn to the messy, multi-dimensional problems in materials science - the ones where experiments and simulations give you mountains of data, but the patterns hide just out of reach. That’s where I found my sweet spot: using machine learning not as a replacement for physical intuition, but as a partner to it.
My work lives at the intersection of real experimental mechanics and computational methods. I’ve spent years in the lab conducting mechanical testing, SEM imaging, and Raman spectroscopy - dealing firsthand with the reality of experimental data: scarce, valuable, and inherently noisy. This experience shaped how I approach ML. When you’ve personally collected 500+ CNT fiber samples and know the effort behind each data point, you learn to extract maximum insight from limited datasets using techniques like LOOCV, data-efficient strategies, and Bayesian optimization that intelligently guides where to experiment next.
My PhD work exemplifies this approach: achieving 200% mechanical property improvements in CNT assemblies by letting algorithms explore parameter spaces while my materials knowledge shaped the constraints. I’ve built physics-informed neural networks that respect governing equations, fine-tuned protein language models, and developed end-to-end MLOps pipelines - each project reinforcing the same principle: the best solutions emerge when computational power meets domain expertise, especially when working with the real, imperfect data that defines engineering practice.
I’m passionate about solving problems where physical intuition meets computational power - where domain knowledge guides algorithms, and algorithms reveal patterns that experiments alone cannot uncover. Every sample I test in the lab informs how I design my ML pipelines. Every model I train reveals new questions to explore experimentally.
I’m looking for roles where I can continue bridging these worlds - where mechanical engineering challenges benefit from ML’s pattern recognition, and ML models are grounded in physical reality. Whether it’s optimizing manufacturing processes, discovering new materials, or building intelligent experimental workflows, I want to work on problems where getting your hands dirty with real data makes all the difference.