What are the responsibilities and job description for the PhD Research Intern, AI position at Stack AV?
Internship Program:
Stack is revolutionizing transportation through AI and is seeking the best and brightest interns to help us realize this vision! As an early-stage start up, we expect our interns to be an integral part of the team, working on robotics research projects that directly impact our product. Along the way, you’ll be provided with opportunities for collaboration and mentorship with industry leaders to accelerate your career and research goals.
We welcome students who are enrolled in university, pursuing a doctorate degree and are currently located in the United States. Summer internships are typically 12 weeks in duration (note: we may be flexible for internships to start in the spring semester or continue into the fall semester).
We offer competitive pay and support sponsorship.
About the Role:
The Artificial Intelligence (AI) Team at Stack develops state-of-the-art solutions for 1) all perception-related jobs, including detection, tracking, mapping, localization, scene understanding, vehicle calibration, and more; and also 2) the data platform and ML infrastructure necessary to succeed as an ML-first autonomous vehicle company.
As an intern in the AI Team, you will have the opportunity to work hand-in-hand with Stack’s world-class researchers and engineers to investigate bleeding-edge problems in AI/CV/ML for autonomous vehicles. In addition to having the opportunity to publish your work, you will learn about the end to end work to build, evaluate, and deploy ML solutions for real robotics applications using multi-modal data, including critical safety systems.
Research Areas:
The project work will be scoped specifically based on your skill set and the research needs of the team, but project areas could include…
- Example project #1: Based on the success of foundational segmentation models such as SegmentAnything (SAM), investigate novel methods for multi-modal semantic segmentation that leverage the growing corpus of labeled image data available to lift it to other sensing modalities.
- Example project #2: Investigate novel methods to perform real-time perception tasks that are suitable for operation during fast relative vehicle motion.
- Example project #3: Investigate temporally-aware data representations in VLMs to answer complex questions about video sequences where ordering is important (e.g: “is there a car going out of turn in a stop sign?”).
- Example project #4: Investigate novel methods to perform panoptic segmentation in multi-modal data sequences with fast relative vehicle motion.
In order to be considered for this team, we strongly encourage students to have the following skills and experiences…
- 3 year PhD student preferred
- Strong foundation in Computer Vision/Machine Learning
- Ability to develop, train, and validate AI/CV/ML models
- Experience with 3D computer vision, multi-modal perception, and/or point cloud processing
- Experience in SOTA deep learning techniques, e.g. multi-modal foundational models
- Strong programming skills in Python or C
- Experience with ML frameworks such as PyTorch
