Internship details:
- Duration: Ideally 12-16 weeks (can extend up to 6 months)
- Level: PhD students (must have prior relevant publications)
- Location: Candidates based in the USA, Canada, or Germany
In this role you will:
- You’ll be part of the ML team tasked with establishing best practises on creating and using synthetic data for perception
- You will work directly with our team to perform cutting-edge research in one or more of these key areas: 3D computer vision, generative models for image/video/scene generation, style transfer, image inpainting, neural rendering, and related areas
- You’ll develop, experiment, and integrate advanced CV algorithms into our platform
- You’ll summarize your findings to help guide our team as well as our customers
- You’ll sync with your team on a daily basis to discuss progress, ideas and problems
- You’ll work with veterans from the gaming, movie and automotive industry
About You:
- You are actively pursuing a PhD in computer science or a related field at a world renowned university.
- You are well-versed with developing machine learning models for computer vision tasks
- You have published peer-reviewed research in at least one of these areas: 3D Computer Vision and Scene Reconstruction, Generative Modeling (GANs, Diffusion models, 3D Gaussian Splatting), Image and Scene Synthesis (Inpainting, Style Transfer), Semantic Segmentation or Automated Scene Understanding, Domain Adaptation or Sim-to-Real techniques
- You have prior experience with Pytorch, Tensorflow, or Keras.
- You're eager to learn You are a team-player and understand the importance of clear communication
- You thrive in ambiguous environments
- You are not afraid to approach new challenges
- You are fluent in english
Bonus Points (Nice to have):
- Experience working with AV simulation, synthetic data generation, or multi-modal sensor data (cameras, LiDAR).
- Familiarity with NeRF, 3D Gaussian Splatting, or point-cloud data.
- Prior experience at a fast-growing startup or dynamic team.
Similar Jobs
What We Do
Training and testing autonomous systems in the real world is a slow, expensive and cumbersome process. Parallel Domain is the smartest way to prepare both your machines and human operators for the real world, while minimizing the time and miles spent there. Connect to the Parallel Domain API and tap into the power of synthetic data to accelerate your autonomous system development.
Parallel Domain works with perception, machine learning, data operations, and simulation teams at autonomous systems companies, from autonomous vehicles to delivery drones. Our platform generates synthetic labeled data sets, simulation worlds, and controllable sensor feeds so they can develop, train, and test their algorithms safely before putting these systems into the real word.
#syntheticdata #autonomy #AI #computervision #AV #ADAS #machinelearning








