10-718: Machine Learning in Practice
A project-based graduate course on building, evaluating, and responsibly deploying machine learning systems — predictive and generative — to solve real-world problems.
Schedule Syllabus Project Grading Policies View on GitHub
Fall 2026
Meets: Tuesday & Thursday, 2:00–3:20 pm, POS 153
People
Instructor
Rayid Ghani (Office: GHC 8023) Office hours: Tuesday 4:00–5:00 pm, Wednesday 4:00–5:00 pm (or email to meet outside these times)
Teaching Assistants
| Olivia Cheng | Namrata Deka | Rachel Kim |
|---|---|---|
| Office hours: TBD | Office hours: TBD | Office hours: TBD |
What this course is
This course gives students experience solving real-world problems with machine learning, working in the gaps between research and practice — where common assumptions (i.i.d., stationarity) break down and where accuracy is only one of the things that matters alongside fairness, robustness, interpretability, uncertainty, and real-world impact. Students work in teams on a semester-long applied project: they identify and scope a real problem, define the analytical formulation and baselines, build models with real data, and evaluate and make recommendations. Generative AI runs as a thread throughout — woven into the core sessions, taught directly in dedicated sessions on foundation models, agents, and generative-system evaluation, and required in the project.
How it’s organized
- Module 1 · Building ML Systems (Weeks 1–6) — the end-to-end ML design and development.
- Module 2 · Beyond the Basic Model (Weeks 9–14) — foundation models & agents, and exploring issues of fairness, robustness, interpretability, uncertainty, and causality.
- Applied ML project — a team-built ML application/system that goes throughout the semester.
Quick links
- Schedule & readings
- Syllabus & learning objectives
- Applied ML project
- Assignments & grading
- Policies & resources
- Role-Playing Seminar Presentations
Previous versions
Fall 2025 · Fall 2023 · Fall 2022 · Fall 2021 · Fall 2020 · Spring 2020