Eindhoven University of Technology (TU/e)
PhD in Communication-Optimized Inference for Large Language Models
- Deadline
- 22 Nov 2026 · closes today Closing soon
- Location
- Eindhoven, Netherlands
- Type
- PhD Position
- Compensation
- EUR 3,204 – 4,051 per month (TU/e PhD scale P)
- Duration
- 48 months · 4 years
- Working language
- English
- Posted
- 21 Nov 2026
- Verified
- 21 Nov 2026
About this opportunity
Eindhoven University of Technology (TU/e) has a four-year, fully paid PhD position in the COIL project. The project models how GPUs communicate over the network when large language models run inference across many machines. You will join the Network Modeling and Design (NoMaD) lab in the Electronic Systems group of the Department of Electrical Engineering, and you will spend about six months at the Technical University of Munich (TUM).
What is offered
- A full-time (1.0 FTE) PhD position for four years, with an intermediate assessment after nine months.
- Research on GPU-to-GPU communication in distributed LLM inference: Network Calculus models, architectural trade-offs, runtime decisions and experiments on programmable network testbeds.
- A research stay of about six months at TUM in Munich.
- Gross salary of €3,204 to €4,051 per month (PhD scale P), plus an 8% holiday allowance and an 8.3% year-end bonus.
- A pension scheme in which TU/e pays 70% of the premiums, 29 to 41 days of leave, training support, on-campus childcare, sports facilities, mental health support, and commuting and home-working allowances.
- International candidates may qualify for the Dutch expat tax scheme.
- Teaching makes up at least 10% and at most 15% of your time each year.
Who can apply
- You need a master's degree in Electrical Engineering, Computer Science, Computer Engineering, Applied Mathematics or a closely related field.
- You should have a strong mathematical background, good programming skills and strong analytical skills.
- You need excellent written and spoken English.
- You must be willing to spend at least six months at TUM.
- Nice to have: experience with computer networks or distributed systems, Network Calculus, simulation, optimization, GPU communication, or programmable networking such as P4 or SmartNICs.
How to apply
- Apply online only, using the Apply button on the vacancy page. Applications sent by email or post are not processed.
- Upload a cover letter, a CV with a publication list and three references, your BSc and MSc transcripts, and an English writing sample on which you are the first author.
- Selection includes interviews and a scientific presentation.
- The closing date is 22 November 2026. The reference number is 2026/677.
- For questions about the project, contact Prof. Lisa Maile (hiring manager). For questions about employment, contact Kevin Caris (HR advisor). Contact details are on the vacancy page.
Requirements
-
Open to all nationalities
-
PhD candidate
-
Minimum degree: Master's
PhD in Communication-Optimized Inference for Large Language Models
22 Nov 2026 · Closes today