PROJECTS
The activities and layout of each project will vary based on content. Average projects will consist of recruitment, member training, and about 3-4 milestones each month.
Join our Discord and ask in the general/help-desk channels!
apply here!
ARCHIVE
Beginner: Not expected to have experience with ML but they should be familiar with programming. Ex. Kaggle competitions, building applications with OpenAI’s API.
Intermediate: Project members have taken CS373/CS471/CS290 or an equivalent class, or self-taught to that level. Project members are familiar with TensorFlow or PyTorch. Ex. Building an autonomous RC car, creating a chatbot from scratch.
Advanced: Project members are going beyond class work knowledge, or the project requires a large time commitment. Ex. Original research (technically difficult), Creating a startup (time commitment).
Fall 2024
BeginnerScraped 10 more years of relevant nfl salary cap, performance, draft, and advanced analytical data and stored them into MongoDB. Created basic UIs for searching based on relevant player data.
Pranay Nandkeolyar
Fall 2024 + Spring 2025
BeginnerLearning through Participation in Kaggle Competitions.
★ Competed in 5+ Kaggle competitions with accuracy rates of upwards of 90%.
Neil Sahai
Fall 2024
BeginnerOver the year, the team was able to successfully develop a fully functional email assistant which allows users to create natural language rules, draft replies using personal context, semantically search an inbox, and handle email using voice. InboxPilot onboarded 15 beta users, processed over 4000 emails, and made over a 100 automatic drafts.
Rishi Mantri
Fall 2024 + Spring 2025
IntermediatePollutant concentration analysis from Satellite Data
Michael J. Wang , Advisor: Prof. Guang Lin, Dr. Gary Doran, Dr. Sina Hasheminassab
Creating an unmanned autonomous boat that can complete complex tasks
★ 2nd Place Trine AIMM ICC competition
Nicholas Wade
Fall 2024
IntermediateAlgorithmic Trading Competitive Team
Yohaan Chokhany
Fall 2024 + Spring 2025
IntermediateThe team successfully generated a synthetic dataset of images taken from in the middle of urban scenes. They used Meta Segment Anything (SAM) model to generate masks, out of which they selected relevant classes, and used the MMSegmentation library to train a new semantic segmentation model.
Manav Gagani
Fall 2024 + Spring 2025
AdvancedDeveloped a novel industrial solution backed by a state-of-the-art learning method in academic research.