Click on each project name for more detailed descriptions.

Industry Projects

**Samba TV: AI/ML for Television — (Lead: Serena Mao HM ‘25, Marwan Bit HM ‘25)**

• SambaTV is a TV data company working on developing Machine-Learning Models to quantify and understand how TVs and associated devices are used, in order to improve user experience.
• To train such models, SambaTV requires large amounts of labeled data (particularly in the form of TV frames with detected logos highlighted regions). We will be utilizing synthetic data generation to generate large swaths of training data.
• We aim to release a small subset of the dataset generated as an open-source project, alongside releasing results regarding how “helpful” synthetic data is for learning.

**Navfeas: Geospatial Data Visualization and Analysis for an Ocean Conservation Startup (Lead: Austin Zang PO ‘24)**

• Learn and practice industry-standard SWE fundamentals: backend API construction, node.js webdev framework, containerization (Docker/Kubernetes), git, unit testing, software architecture, databases (MongoDB, Neo4j)
• Work with geospatial GPS coordinates to predict whale migration patterns - aiding mariners to avoid areas where shipping traffic creates an environmental disturbance

Student Projects

**p-okémon: Pokémon Battle Engine (Lead: Zac Davis CM ‘24)**

• Develop an evaluation function that predicts a win% in a given position using VGC statistics APIs and battle simulators.
• Create models using supervised learning techniques to iterate on a search tree of potential outcomes and select the strongest outcomes determined by an evaluation function.
• Employ self-play training to further improve the supervised models and launch unsupervised models which solely learn through self-play.