ECE 410/510: Hardware for AI and ML
News and updates:
- Feb 12, 2026: Course description, learning outcomes, course schedule, and FAQs were updated.
Course description:
This course covers the design, simulation, optimization, and evaluation of specialized hardware for artificial intelligence and machine learning workloads. Students begin with the mathematical foundations that connect AI/ML algorithms to their hardware implementations, then apply HW/SW co-design methods to map algorithms such as CNNs, DNNs, and transformer-based LLMs onto architectures including GPUs, TPUs, FPGAs, systolic arrays, and neuromorphic processors. Additional topics include computational profiling with Python and CUDA, benchmarking across hardware platforms, in-memory computing with memristive crossbar arrays, and the use of LLMs as hardware design tools for HDL generation and physical design automation. The course alternates between lectures and hands-on codefest sessions where students tackle open-ended design challenges, from implementing neurons and systolic sorting to running simulations on neuromorphic hardware platforms such as BrainScaleS-2.
Course organization:
The course follows a weekly cycle: Monday lectures introduce concepts, methods, and context, while Wednesday codefest sessions are hands-on working sessions where students tackle open-ended design challenges that apply and extend the lecture material. Codefests are collaborative and may involve programming, simulation, hardware design, benchmarking, and documentation. Students are expected to use LLMs and other modern tools during codefests.
Learning outcomes:
After successfully completing this course, students will be able to:
- Apply the mathematical foundations underlying AI/ML algorithms (linear algebra, calculus, probability) in the context of hardware design and optimization.
- Explain the principles and tools for SW/HW co-design.
- Explain the foundations of neural networks and Large Language Models (LLMs), including CNNs, DNNs, and transformer architectures.
- Explain the foundations of specialized hardware for AI/ML, including GPUs, TPUs, FPGAs, systolic arrays, and neuromorphic architectures.
- Profile, benchmark, and analyze the performance of AI/ML workloads across different software frameworks and hardware platforms.
- Map AI/ML algorithms onto specialized hardware through co-design methodologies.
- Evaluate and optimize HW designs through co-design for computational power, energy efficiency, and flexibility.
- Use modern software and hardware tools, including LLM-assisted design flows, for designing and deploying specialized AI/ML hardware.
- Explain emerging computing paradigms, including in-memory computing, memristive devices, and their applications to AI/ML workloads.
Tentative course plan:
General catalog and banner information:
- Course prefix: ECE
- Course number: 410/510
- Catalog course title: Hardware for Artificial Intelligence and Machine Learning
- Credit hours: 4
- Grading: Letter grade
- Course intended for: Graduate and undergraduate students
- Instructional method: Lecture (Mon) + codefests (Wed)
- Prerequisites (recommendations, will not be enforced):
- Undergraduate: ECE 371 and ECE 351
- Graduate: ECE 485
FAQs:
- Q: What’s the course structure? A: The course follows a weekly cycle: Monday lectures introduce concepts, methods, and context, while Wednesday codefest sessions are hands-on working sessions where students tackle open-ended design challenges that apply and extend the lecture material. Codefests are collaborative and may involve programming, simulation, hardware design, benchmarking, and documentation. Students are expected to use LLMs and other modern tools during codefests.
- Q: How can I prepare for this course? A: Strengthen your Python programming skills. Modern hardware engineering increasingly relies on software proficiency. Whether you’re designing circuits, architecting systems, or developing algorithms, strong programming fundamentals are essential.
- Q: What will be the balance between theory, foundations, and practice? A: The course prioritizes foundations and theory over trendy tools and frameworks. AI hardware evolves rapidly; specific tools you learn today will likely be superseded by semester’s end. What endures, and what best serves your long-term career, are the underlying principles and theoretical frameworks.
- Q: How is the course aligned with current industry trends? A: We engage with industry developments where pedagogically valuable, but trends shift weekly. Your career will span decades and multiple technology cycles. Time invested in mastering fundamentals yields far greater returns than chasing ephemeral trends.
- Q: Will the prerequisites be enforced? A: No. Prerequisites serve as recommendations and cannot be enforced for 400/500-level courses. If you believe your background prepares you for success here, you’re welcome to enroll. However, recognize that you assume responsibility for any gaps in preparation.
- Q: What projects will you assign? A: This is a very project-based class. Projects will be introduced during class sessions and tend toward open-ended, challenging problems. They provide opportunities for both learning and demonstrating capability. There will be one large final project that starts during week 1. The weekly projects and codefests are building blocks for the final project.
- Q: Will projects be collaborative? A: Projects are individual assignments. This structure allows you to develop mastery independently and provides a clearer picture of your individual capabilities and growth.
- Q: Will this class be recorded? A: The course is offered in-person only, with no lecture recordings. Active participation in both lectures and codefests is essential, as much of the learning happens through real-time problem-solving, discussion, and collaboration that cannot be replicated asynchronously. If you cannot attend regularly or prefer remote learning, consider alternative courses that better fit your needs.
- Q: Will this be an easy class? A: Engaging with cutting-edge technology at the frontiers of innovation requires substantial effort and persistence. If you’re seeking a lighter workload, other courses may be more appropriate.
- Q: Will AI tools be allowed in this course? A: AI tools are not merely allowed but expected. Success in this course will likely require a professional-tier subscription to your preferred AI platform (ChatGPT Plus, Claude Pro, Gemini Advanced, or similar).
- Q: Can I take this class if I am opposed to using AI? A: You’re welcome to enroll, but assignments are designed with AI integration in mind. Completing them without these tools will be significantly more difficult and may not be practical.