Software Engineer, Perception, Scene Understanding

The Perception Scene Semantics team is responsible for developing ML models for scene understanding around construction, emergency scenes, mapless driving and world changes. The team works on the end to end ML pipeline from data mining, labeling, training onboard / offboard models, and deployment. We need someone with strong ML experience to drive the architecture upgrades and product feature development.

In this hybrid role, you will report to a Technical Lead Manager within Semantics.

You will:

• Develop and deploy deep learning models, including vision language models (VLMs) and Large Language Models (LLMs), multi-modal sensor fusion architectures for scene understanding
• Develop data mining, labeling, training and eval pipelines to support the onboard development
• Collaborate and work in partnership with research teams across Alphabet
• Design large-scale foundation models trained on our vast data

You have:

• BS degree in Computer Science or a similar discipline, or an equivalent amount of deep learning experience
• 3+ years of industry or research experience with Large Language Models (LLM), Vision Language Models (VLM) or Computer Vision
• 3+ years experience with Python
• Hands-on experience with deep learning frameworks such as Jax, PyTorch, Tensorflow

We prefer:

• PhD degree in Computer Science or a similar discipline, or an equivalent amount of deep learning experience
• Publications on top-tier conferences like CVPR/ICCV/ECCV/ICLR/ICML/NeurIPS/ICLR/AAAI/IJCV/PAMI
• Github repositories or Tech Blogs of LLMs/ VLMs
• Experience with C++

#LI-hybrid