Machine Learning
Work With Us!
We're eager to partner with organizations, researchers, and individuals who want to enhance ML education. From client projects to educational ideas and outreach, your support helps our community grow!
Inclusive machine learning community powered by Duke talent.
Duke Applied Machine Learning (DAML) is a student-led organization that pairs education, research, and client delivery. We help curious students become consultants who can design and deploy ML systems, lead projects, and collaborate with partners across Duke and beyond.
Programs that cultivate machine learning talent at Duke.
An eight-week ML fundamentals course where new members learn Data Science and program their first ML projects.
Hands-on deep dives into deploying ML solutions through containerization, CI/CD pipelines, and more into an UI platform.
Office hours pair first-years wtih our experienced members.
Client projects —
You get Duke's CS Talent,
We get real-world experience.
Our team of Product Managers and top Division Leads are happy to meet with partners to discuss ML projects: incorporating KPIs and ML expertise to your business.
We match a team of our Engineers to each partner for client projects. Partners get top Duke CS Talent, and our members get real-world experience.
Our consultants provide an in-depth project plan, DAML teams provide data analysis reports, model prototypes & reports on performance, and deploys models along with user guides.
Members ship research, software, and strategy that matter.
Explore a snapshot of our member's projects from recent semesters.
Spring 2024
Jai Kasera
Retrieval-augmented chatbot that surfaces the latest Senate bills, hearings, and votes by scraping and indexing US Congress data for precise policy answers.
View repo ->Spring 2025
Haiyan Wang, Benjamin Yan, Jai Kasera
PyTorch engine inspired by AlphaZero that learns entirely via self-play using deep reinforcement learning and Monte Carlo Tree Search.
View repo ->Spring 2024
Sam Borremans, Samuel Orellana Mateo, Yash Singam, Samir Travers, Benjamin Yan
Compared CNN architectures like ResNet, EfficientNet, and ShuffleNet while mitigating background bias in the Maize Leaf Disease dataset.
View repo ->Fall 2023
Brian Chen, Arthur Zhao, Darian Salehi, Jai Kasera
LSTM and BERT-driven classifier that flags toxic Twitter content, supporting safer communities through high-accuracy moderation tooling.
Summer 2025
Kevin Mao
FastF1, XGBoost, and SHAP-powered model predicting whether a driver will finish ahead of their grid start using weather, pit strategies, and historical performance.
View repo ->Build alongside Duke's most curious engineers
Join a community that ships ML projects, runs member-led labs, and mentors across research and production. We welcome members at all experience levels — start small, learn quickly, and lead real work.