NUS - Wei Li is Hiring

I am an Assistant Professor in the School of Computing at the National University of Singapore (NUS). My group builds intelligent design automation and testing systems that make chip development faster and more reliable. I am looking for Ph.D. students, postdocs, and master's, undergraduate, and visiting researchers who are excited about EDA, hardware testing, AI/LLMs, and future hardware.

AI is driving unprecedented demand for computing, and hardware must learn to answer faster.

Ph.D. from CMU, B.Sc. and M.Phil. from CUHK · 2026 ML and Systems Rising Star · Best Paper Awards at ASP-DAC, ISSTA, and ICTAI · More about me

Openings

Ph.D.Open

Email me early. Admission is committee-based, and I actively support strong-fit applications through NUS Computing.

  • 15 Dec 2026→ Aug 2027 intake
  • 15 Jun 2027→ Jan 2028 intake
Postdoc: AI for Chip Design1 position

2–3 years, co-advised with Prof. Tulika Mitra, Dean of NUS School of Computing.

Good fit for: AI/ML for EDA, LLM agents for chip design, hardware/software co-design.

Postdoc Fellowship: 3D-IC Physical DesignOpen

18–24 months, through the NUS ARTIC Fellowship. Physical design for multi-tier 3D ICs, e.g., 3D DRAM processing-in-memory.

Requires: a Ph.D. obtained within six years of the start date. Fellows receive their own research grant.

  • 8 Nov 2026→ application deadline; email me as soon as possible
Master's / Undergrad / VisitingOpen

Email me with your CV and interests (see How to Apply).

Why Now

Turning a new architecture into silicon takes over two years and over 1000 engineers
Turning a brilliant new architecture into competitive silicon can still take two years and 1000+ engineers, while AI models change every few months. That gap is what my group attacks. Read the full research vision →

Research Directions

Candidate routes converging to one optimized route

1. EDA Algorithms: Optimization & AI

Differentiable optimization, reinforcement learning, graph neural networks, ...

New algorithms for core EDA problems such as routing and layout decomposition.

Good fit if you like: algorithms, optimization, machine learning, CS theory.

Example: differentiable global routing (DAC'24).

Read more →

An LLM agent driving the chip design flow

2. Agentic EDA

How do we bring AI agents into EDA?

Agents that read RTL, layouts, timing reports, and tool logs, call EDA tools, check their own progress, and know when to ask a human.

Good fit if you like: LLM agents, reasoning, tool use, multimodal learning.

Example: BRIDGES, connecting graph-structured chip data with LLM reasoning.

Read more →

A faulty core caught by an in-field test sweep

3. Hardware Testing

Can we trust chips after fabrication?

Some defective chips pass manufacturing test and later produce wrong results in datacenters without warning (silent data corruption). We study why defects escape and how to catch them, before and after deployment.

Good fit if you like: testing, reliability, systems, working with real industrial data.

Example: ITC'22: the standard industrial fault model matched only ~5% of thousands of real failures.

Read more →

A monolithic 3D stack and a photonic ring resonator

4. Future Hardware

Design automation for 3D ICs, memory systems, photonics, ...

As 2D silicon scaling slows, new hardware needs its own design and test tools. We build them together with device and architecture researchers.

Good fit if you like: architecture, emerging devices, cross-disciplinary work.

Read more →

What You Get

  • Industry: I interned in design-automation research teams at Apple and NVIDIA, held the Apple Ph.D. Fellowship (twice) and the Qualcomm Innovation Fellowship, and stay close to EDA/testing teams and AI-for-chip startups. I strongly encourage internships and will actively help you get them.
  • Academia: I trained at CMU and CUHK, and work with collaborators across EDA, testing, and AI. I will help you build taste, publish strong work, and meet the right people.
  • Funding: The NUS Research Scholarship covers the first four years. If a fifth year is needed, finding funding is my job, not yours.
  • Advising: Weekly meetings, usually two active projects, and no rigid paper count. Read my advising statement →

Who Should Apply

You do not need an EDA or hardware background. Many chip-design problems are optimization, graph, generative-modeling, reinforcement-learning, or agentic problems in disguise. My own undergraduate major was computer science. What matters most is curiosity, discipline, and the courage to work on problems that are larger than one person.

Coming from AI/CS? Here is a starter guide and reading path →

How to Apply

Email weili3@nus.edu.sg with:

  • your CV, and your transcript if available;
  • a short description of your research interests;
  • one or two projects or papers that made you curious.

Please use the subject line [Position, Intake] Your Name - Direction, for example [PhD, Aug 2027] Jane Doe - Direction 2.

We are still early. The map is not finished (that is exactly why it is a good time to join).