Multimodal Sensing AI Research Intern

Bosch; Pittsburgh (Allegheny County); On-site; USD 33,000 - 47,000; Part time

Job summary

BOSCH is seeking a Multi-modal Sensing AI Research Intern in Pittsburgh to develop state-of-the-art multi-modal models for active (radar, ultrasound) and passive sensing (acoustic, vibration, EEG) using signal processing and machine/deep learning.

Ideal candidates are PhD students in CS/EE with 2+ years of programming experience and proficiency in PyTorch Lightning and HuggingFace Transformers. You will collaborate with researchers and summarize findings for publications or patents.

Qualifications

• Currently enrolled as PhD student in Computer Science, Electrical Engineering, or related fields.
• 2+ years programming experience, proficiency in PyTorch (Lightning), HuggingFace (transformers), hydra.
• Broad knowledge of machine- and deep-learning algorithms and principles and state-of-the-art methods.
• Minimum GPA of 3.0

Responsibilities

• Develop state-of-the-art multi-modal models for active (radar, ultrasound) and passive sensing (acoustic, vibration, EEG) use-cases using combination of classical signal processing and machine/deep learning-based approaches.
• Research and develop solutions for multi-modal representation learning and modality adaptation with paired / unpaired sensor data.
• Collaborate with other researchers to evaluate the developed model on downstream applications.
• Summarize research findings in high-quality papers and/or patent submissions.

Skills

Python programming

Machine learning

Deep learning

Research mindset

Education

PhD student in Computer Science / Electrical Engineering or related fields

Tools

PyTorch Lightning

HuggingFace Transformers

Hydra

Job description

Your tasks

As the Multi-modal Sensing AI Research Intern, a few of your key responsibilities will include:

• Develop state-of-the-art multi-modal models for active (radar, ultrasound) and passive sensing (acoustic, vibration, EEG) use-cases using combination of classical signal processing and machine/deep learning-based approaches.
• Research and develop solutions for multi-modal representation learning and modality adaptation with paired / unpaired sensor data.
• Collaborate with other researchers to evaluate the developed model on downstream applications.
• Summarize research findings in high-quality paper and/or patent submissions.

Your profile

Minimum Qualifications:

• Currently enrolled as PhD student in Computer Science, Electrical Engineering, or related fields.
• 2+ years programming experience, proficiency in PyTorch (Lightning), HuggingFace (transformers), hydra.
• Broad knowledge of machine- and deep-learning algorithms and principles and state-of-the-art methods.
• Minimum GPA of 3.0

Preferred Qualifications:

• Experience using raw sensor data (eg. Radar, ultrasound, acoustic etc.) in machine learning projects.
• Knowledge of digital signal processing principle and methods, multimodal representation learning.
• Publication record in top machine learning and signal processing venues.
• Experience with HPC platforms and job managers (Slurm, IBM LSF).

Contact & additional information

Equal Opportunity Employer, including disability / veterans