Mitsubishi Electric Research Laboratories
Internship - Radar/Sensor Reasoning Models
At a glance
- Term
- Starts May 2027
- Length
- 13 weeks
- Preferred skills
- generative AI & LLMsMachine learningPythonPyTorch
Listing details
About the job
The Computation Sensing team at MERL is seeking a highly motivated intern to conduct fundamental research on sensor reasoning models—algorithms that can understand, explain, and act on multi-sensor data (e.g., radar, infrared, LiDAR, event camera) through text-, visual-, and multimodal reasoning. Ideal candidates will be comfortable bridging modern perception (detection/segmentation/tracking) with higher-level reasoning capabilities. Experience with text, visual, and multimodal reasoning is highly preferred. The intern will work closely with MERL researchers to develop novel algorithms, design experiments using MERL’s in-house testbeds, and prepare results for patents and publication. The internship is expected to last 3 months, with a flexible start date from May 2027.
Required Specific Experience
- Reasoning with sensor data: Demonstrated work in text-, visual-, and multimodal reasoning (e.g., VQA over sensor streams, temporal/spatio-temporal reasoning, chain-of-thought, instruction following).
- LLMs & VLMs for sensor perception: Experience aligning or conditioning LLMs/VLMs on sensor outputs (e.g., point clouds, radar heatmaps, BEV features).
- Perception foundations: Solid understanding of state-of-the-art transformer-based and diffusion-based frameworks
- Datasets & evaluation: Hands-on experience with open large-scale multi-sensor datasets (e.g., OctoNet, CUHK-X) and open radar datasets (e.g., MMVR, HuPR, HIBER, RT-Pose, K-Radar). Ability to design reasoning-centric benchmarks.
- Proficiency in Python and deep learning frameworks (PyTorch/JAX), plus experience with GPU cluster job scheduling and scalable data pipelines.
- Proven publication record in top-tier venues such as CVPR, ICCV, ECCV, NeurIPS, ICLR, ICML (or equivalent).
- Knowledge of sensor (RF, infrared, LiDAR, event camera) fundamentals; for radar, familiarity with FMCW, MIMO, Doppler signatures, radar point clouds/heatmaps, and raw ADC waveforms.
- Familiarity with MERL’s recent radar perception research, e.g., TempoRadar, SIRA, MMVR, RETR, RAPTR, REXO.
The pay range for this internship position will be 6-8K per month.
Research Areas
Artificial Intelligence, Computational Sensing, Machine Learning, Signal Processing
Host
Perry Wang