We are pleased to announce that the Machine Learning Summer School 2026 will run for two weeks in June 2026 in New York City, at the Columbia University campus. We will host approximately 200 PhD students alongside key faculty, industry speakers, and invited practitioners to take part in a rigorous program that will balance practical training on state-of-the-art systems (evaluation, agentic AI, RAG, data pipelines) with forward-looking research areas (alignment/safety, interpretability, verification & reasoning).

Our objectives are to deliver a rigorous curriculum, an excellent experience for participants, and measurable impact on research. The program will combine lectures, tutorials, invited talks, and hands-on labs. In addition to reinforcement learning theory, LLM alignment/safety, RAG & agents, and time series analysis, the program will include systems and efficiency for LLMs, post-training and preference optimization, synthetic data practices, evaluation, mechanistic interpretability, and reasoning.

The event is being jointly overseen by a Steering Committee and a Local Organizing Committee led by Columbia University for campus operations. The organizers represent Bloomberg, Columbia Engineering (SEAS) and Columbia’s Data Science Institute (DSI), the NYU Center for Data Science (CDS), and Cornell Tech.

At a glance:

• When: June 15-26, 2026
• Where: Columbia University Morningside campus
• Who: ~200 PhD students
• Academic partners: Columbia University, NYU CDS, Cornell Tech

Organizers & Steering Committee: 

• Ali Hirsa (Columbia),
• Gary Kazantsev (Bloomberg/Columbia),
• David Rosenberg (Bloomberg/NYU),
• Alex Smola (Boson.ai),
• Paola Cascante-Bonilla (Stony Brook University),
• Carlos Fernandez-Granda (NYU CDS),
• Andrew Owens (Cornell Tech)