About the role#
Bosch Research and Technology in Pittsburgh is looking for a PhD student to join the team as a Research Intern. You will focus on advancing foundation models for multimodal sensing. The work explores how models learn from diverse sensor streams to understand physical environments, recognize activities, and transfer knowledge across different sensing systems. You will work with experienced researchers to develop methods at the intersection of deep learning, signal processing, and cyber-physical systems.
What you'll need#
- Currently pursuing a PhD in Computer Science, Electrical Engineering, Computer Engineering, Robotics, Data Science, or a related field.
- Strong knowledge of machine learning and deep learning principles.
- Proficiency in Python and experience with frameworks like PyTorch or TensorFlow.
- Experience with multimodal learning, time-series modeling, computer vision, or sensor data analysis.
- Strong problem-solving and communication skills.
- Ability to work effectively in a research environment.
- Interest in developing transferable learning methods for real-world sensing.
- Research experience in foundation models, self-supervised learning, contrastive learning, masked modeling, or generative modeling is preferred.
- Experience with raw sensor data, signal processing, sensor fusion, domain adaptation, or few-shot learning is a plus.
- Familiarity with efficient model architectures, edge AI, or embedded systems is helpful.
- Publications in machine learning, computer vision, robotics, or signal processing venues are preferred.
What you'll do#
- Design, train, and evaluate foundation models for heterogeneous sensor data including radar, acoustic, vibration, image, lidar, Wi-Fi, and inertial signals.
- Investigate self-supervised learning, multimodal representation learning, and cross-modal alignment using both labeled and unlabeled data.
- Create methods for sensor fusion and adaptation across tasks, environments, and device configurations.
- Evaluate models on tasks like scene understanding, activity recognition, event detection, tracking, and anomaly detection.
- Explore model robustness to missing modalities, noisy measurements, and changing operating conditions.
- Collaborate with cross-functional teams to prototype sensing solutions.
- Document and share findings through reports, presentations, and potential publications or patents.
Location & details#
- Location: Pittsburgh, Pennsylvania, United States.
- Modality: On-site.
- Employment type: Full-time intern.
- Term: Rolling.
- Requirements: Indefinite U.S. work authorization is required. Employment is contingent upon passing a drug screen and background check.


