PhD Position in Foundation Models for Automotive Imaging Radar, TU Delft (with NXP and Perciv.AI)

Job Description

TU Delft has a PhD position in foundation models for radar perception in cars. You will learn general radar representations from mostly unlabelled data that work across different radar sensors and tasks. The project is with Perciv.AI and NXP Semiconductors, in the Intelligent Vehicles group.

At a glance

• Position: PhD candidate, 4 years (TU Delft job 3888)
• Where: Intelligent Vehicles group, Department of Cognitive Robotics, Faculty of Mechanical Engineering, TU Delft
• Pay: €3,204 to €4,051 gross a month (year 1 to year 4), plus 8% holiday allowance and an 8.3% year-end bonus
• Supervisors: Prof. Dariu Gavrila and Dr. Julian Kooij
• Apply: 18 October 2026

What the radar foundation model PhD works on

• Learn general radar representations from largely unlabelled data
• Adapt them to perception tasks such as 3D object detection and free-space estimation with little labelled data
• Transfer knowledge from vision and LiDAR foundation models to radar, and build multimodal models that include radar
• Study how one model can handle different radar representations and sensor setups
• Possibly model dynamic radar scenes with generative or predictive methods
• Test on real data from research vehicles and advanced radar prototypes

Why this role matters

Imaging radar works in rain, fog and darkness, where cameras and LiDAR struggle, and it is cheap enough for mass-market cars. But today’s radar networks are built for one sensor and one task, and labelled radar data is scarce. A radar foundation model would fix both. NXP makes automotive radar chips, so the work links AI research to real semiconductor products. The project is part of the NWO-funded FIND programme.

What TU Delft is looking for

• An MSc in computer science, artificial intelligence, robotics, electrical engineering or a closely related field
• A strong academic record and a solid background in machine learning and deep learning
• Skill in developing and critically evaluating ML research software, preferably in Python and PyTorch
• Interest in foundation models, self-supervised, multimodal learning and 3D perception
• Good written and spoken English

Nice to have

• Experience with radar, signal processing, computer vision, autonomous driving or sensor fusion

Pay and location

Pay follows the Collective Labour Agreement for Dutch Universities: €3,204 gross a month in the first year, rising to €4,051 in the fourth, for a 38-hour week. There is an 8% holiday allowance and an 8.3% year-end bonus. The four years come as a 1.5-year contract with a go/no-go review within 15 months, then a 2.5-year contract. TU Delft runs a knowledge-security risk assessment in the final stage of selection.

How to apply

Apply through the official TU Delft page. The closing date is 18 October 2026. Apply online with a motivation letter, a detailed CV, BSc and MSc transcripts, and any supporting material such as your MSc thesis, publications or code. References are asked for only if you are invited to interview. Email applications are not processed.

See all our PhD positions, or browse more research jobs on ResearchJobs.in.