Lili Karashchuk
Artificial Intelligence Researcher at Howard Hughes Medical Institute (HHMI)
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolLili Karashchuk is an Artificial Intelligence Researcher with 13 years of experience bridging neuroscience and machine learning, currently based in Seattle and working at the Howard Hughes Medical Institute. She holds a PhD in Neuroscience from the University of Washington and a BA in Statistics and Computer Science from UC Berkeley, grounding her work in both experimental brain science and rigorous quantitative methods. Lili has led scientific teams as Chief Science Officer and contributed at research institutes like the Allen Institute for Neural Dynamics, translating neural dynamics insights into scalable AI tools. An active open-source contributor, she improved DeepLabCut’s pose-estimation pipeline to support multiple maxima per bodypart and more robust batch/video prediction handling—work that impacts a widely used tool for markerless animal pose tracking. Her background blends hands-on ML engineering with experimental design, enabling her to prototype models that are both biologically informed and production-ready. Colleagues describe her as a “neuroscience witch,” reflecting a knack for intuitive, creative problem-solving that uncovers nonobvious patterns in complex data.
13 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy - PhD Neuroscience, Doctor of Philosophy - PhD Neuroscience at University of WashingtonBachelor of Arts (B.A.) Statistics and Computer Science, Bachelor of Arts (B.A.) Statistics and Computer Science at University of California, Berkeley
English, French, Russian
Github Skills (9)
computer-vision10
machine-learning10
deep-learning10
tensorflow10
python10
pose-estimation10
numpy9
feature-detection8
keypoint8
Programming languages (7)
Github contributions (5)
DeepLabCut/DeepLabCut
Jan 2019 - Dec 2019
Official implementation of DeepLabCut: Markerless pose estimation of user-defined features with deep learning for all animals incl. humans
Role in this project:
ML Engineer
ML EngineerContributions:6 commits, 5 PRs, 13 comments in 10 months
Contributions summary:Lili primarily contributed to the `deeplabcut` repository by modifying the pose estimation code, specifically the `predict_videos.py` and `nnet/predict.py` files. Their contributions include adding functionality to return multiple maximums per bodypart, correcting a typo, updating the code to handle different output configurations, and improving prediction-related operations. The user also made several updates related to progress bar handling and batch size configuration within the video analysis pipeline.
pytorchkeypoint-detectiondefinedmarkerlessdeep-learning
lambdaloop/aniposelib
Jun 2019 - Nov 2021
Contributions:139 commits, 9 PRs, 98 pushes in 2 years 5 months
cameraimage-processingpythoncameras
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.