Lisa Dunlap
PhD Student at University of California, Berkeley
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolLisa Dunlap is a Data Scientist with eight years of experience who brings a research-first approach to applied machine learning. She holds a PhD-level research background at UC Berkeley focused on vision & language and automated data science, and collaborates with groups like UC Brise and BerkeleyVL. Lisa bridges research and engineering—contributing to open-source work such as neural-backed-decision-trees where she developed detailed CIFAR10 analyses and visualizations (per-class accuracy, path length, backtrack counts) that support making decision trees competitive with neural nets. Her strengths lie in data exploration, preprocessing, model interpretability, and turning complex metrics into actionable insights. Known for uncovering non-obvious class-level failure modes, she applies meticulous analysis to improve model reliability in production contexts.
9 years of coding experience
5 years of employment as a software developer
Bachelor of Arts - BA, Mathematics and Computer Science, Bachelor of Arts - BA, Mathematics and Computer Science at University of California, Berkeley
Github Skills (9)
pandas10
decision-tree10
jupyter-notebook10
seaborn10
cifar10010
python10
matplotlib10
image-classification9
interpretation9
Programming languages (3)
Github contributions (5)
alvinwan/neural-backed-decision-trees
Feb 2020 - Feb 2020
Making decision trees competitive with neural networks on CIFAR10, CIFAR100, TinyImagenet200, Imagenet
Role in this project:
Data Scientist
Data ScientistContributions:11 commits in 6 days
Contributions summary:Lisa's contributions center on analyzing and visualizing CIFAR10 decision tree metrics. They added a Jupyter Notebook to analyze the CIFAR10 dataset, including preprocessing steps like loading and parsing data. The notebook calculates and presents per-class statistics such as accuracy, path length, and backtrack counts. Subsequent commits merged visualization-related changes, indicating a focus on data exploration and analysis.
pytorchdecision-makingimagenettreesresnet
lisadunlap/VAE-GAN
May 2019 - Dec 2019
Contributions:6 commits, 9 pushes, 1 branch in 6 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.