Course Syllabus

Table of contents

1. Reading
2. Class Format and Tentative Schedule
3. Class Participation
4. Presentation
5. Late policy
6. Grading
7. Academic Integrity
8. Students with disabilities or learning needs

Reading

Required readings are from the open-source textbook Machine Learning Systems:

• Vol1: Introduction to Machine Learning Systems. The MIT Press (2026). Vijay Janapa Reddi. (free book).

• Vol2: Machine Learning Systems at Scale. The MIT Press (2026). Vijay Janapa Reddi.
  (free book).

Class Format and Tentative Schedule

The course consists of approximately two-thirds instructor-led lectures during the first ten weeks and one-third student-led paper presentations and guest lectures during the final five weeks.

• Week 1-3: AI systems foundation
• Week 4: AI compute infrastructure
• Week 5: AI network infrastructure
• Week 6: AI storage infrastructure
• Week 7-8: Distributed training and performance engineering (FlashAttention)
• Week 9-10: Inference at scale (vLLM, KVC, speculative decoding)
• Week 11-14: Student presentations and guest lectures on selected AI systems papers, including production-scale training systems, the DeepSpeed system series, KV-cache optimization, etc.
• Week 15-16: Final project showcase

Class Participation

• Attendance: Regular attendance is expected and will be taken into consideration when grading class participation.

• Engagement: You are expected to actively participate in class discussions in a quality way.

  • By “quality”, we mean that you’re contributing meaningfully to the discussion (e.g., highlighting genuine confusions, expanding idea to other contexts, answering leader’s questions, providing new angles of seeing the paper), instead of distracting the discussion (e.g., ask irrelevant questions, add superficial comments, echo other people).

Presentation

Each student will be assigned one or more research papers (based on class size) to present during Week 10 to Week 14.

• Preparation: Prepare a 25-30 minutes presentation on the paper, summarizing the research problem, methods, results, and implications. Please send the slides to the instructor at least 3 days before the presentation.

• Discussion: Be prepared to answer questions and lead a discussion about the paper after your presentation.

Late policy

Late policy applies differently to review submissions and project submissions.

• For survey submissions: We do not accept late submissions as you already get 5 wild cards.

• For assignment and project submissions: Your work is late if it is not turned in by the deadline.

  • 10% will be deducted for late submissions each day after the due date. That is, if an assignment is late, we’ll grade it and scale the score by 0.9 if it is up to one day late, by 0.8 if it is up to two days late, and by 0.7 if it is up to three days late.
  • Late submissions will only be accepted for 3 days after the due date. Assignments submitted more than 3 days late will receive a zero. If you’re worried about being busy around the time of a HW submission, please plan ahead and get started early.
  • Assignment that does not compile or run will receive at most 50% credit.

IMPORTANT: Debugging distributed systems can be time-consuming, even with AI’s help. You need experience to harness AI to make things right. So, please plan ahead and get started early!

For fairness to all, there are no exceptions to the above rules.

Grading

Your grade will be calculated as follows:

• Assignments: 30%
• Class participation: 10%
• Paper presentation: 10%
• Project: 50%

Academic Integrity

The School relies upon and cherishes its community of trust. We firmly endorse, uphold, and embrace the University’s Honor principle that students will not lie, cheat, or steal, nor shall they tolerate those who do. We recognize that even one honor infraction can destroy an exemplary reputation that has taken years to build. Acting in a manner consistent with the principles of honor will benefit every member of the community both while enrolled in the School of Data Science and in the future. Students are expected to be familiar with the university honor code, including the section on academic fraud.

Students with disabilities or learning needs

It is my goal to create a learning experience that is as accessible as possible. If you anticipate any issues related to the format, materials, or requirements of this course, please meet with me outside of class so we can explore potential options. Students with disabilities may also wish to work with the Student Disability Access Center to discuss a range of options to removing barriers in this course, including official accommodations. Please visit their website for information on this process and to apply for services online. If you have already been approved for accommodations through SDAC, please send me your accommodation letter and meet with me so we can develop an implementation plan together.