Aspiring Research Scientist · Data Scientist · AI/ML
I study how AI systems behave when the stakes are human: when should people trust AI, and how do we measure that honestly? My work sits in responsible AI and LLM evaluation. I hold a Master's from Purdue University 🎓 and have co-authored peer-reviewed research in applied machine learning.
My path hasn't been a straight line. After a hard reset, I rebuilt in under a year and found the question I want to spend my career on: how trustworthy AI is when it makes judgments about people, and how to tell what the person contributes versus the machine.
Currently building toward trustworthy AI evaluation: benchmarks drawn from real human language rather than simulated data, careful comparison of LLM judgments with expert human judgment, and privacy-first data practices. Most models are tested on data where the answer is known by design; I want to know how they hold up in the messy real world.
When people use AI advice, who benefits and who gets misled? A study of how people's own skill, and the way AI is presented to them, shape whether they follow good advice or bad advice. Write-up coming soon.
Isolating human skill from machine performance. Using 64,000 IMSA endurance-racing laps, I compared teammates who shared the same car, so the car is removed from the comparison by design.
- Gold drivers look 0.17 s/lap slower than Platinum across different cars, but only 0.03 s/lap (not significant) against a teammate in the same car: most of the apparent skill gap is the machine.
- The same "hold the machine fixed, measure the person" logic underpins how AI tools and self-driving cars should be judged.
mixed-effects modelsbootstrapcausal inferencePython
Helps people find the files they're paying to store but don't need, across phone, laptop and cloud, built as human-centered AI where people stay in control of what gets deleted.
With a Ph.D. research group at Virginia Tech, I:
- Co-authored papers on predicting mental health (preprint), opioid crisis detection (preprint) and hotel industry trends (paper) using machine learning.
- Received recognition at ASONAM 2023 (IEEE/ACM Advances in Social Networks Analysis and Mining), Industrial Track. 📚
Strengths: statistical modelling and causal inference · experimental design and evaluation · machine learning · NLP · data cleaning and visualization
Research Scientist · Applied Scientist · Data Scientist · Machine Learning Engineer · Responsible AI / AI Evaluation roles