Find your next role

Description

Amazon's Last Mile Geospatial Science team leverages sate of the art computer vision, generative AI, and deep learning to enhance vehicle navigation and ensure safe, efficient deliveries by enriching map data from billions of satellite, aerial, and street-level images and videos. The role involves building large-scale machine learning systems that analyze terabytes of multimodal data to solve novel problems, translating business requirements into prototypes while prioritizing driver and customer safety. The team seeks scientists who can combine domain expertise with machine learning to invent and implement state-of-the-art solutions in a collaborative environment with direct business impact.

Key job responsibilities
Successful candidates should have a deep knowledge (both theoretical and practical) of various machine learning algorithms for large scale computer vision problems, the ability to map models into production-worthy code, the communication skills necessary to explain complex technical approaches to a variety of stakeholders and customers, and the excitement to take iterative approaches to tackle big, long term problems. The applied scientist should be proficient with image and video analysis using machine learning, including designing architecture from scratch, modify existing loss functions, full model training, fine-tuning, and evaluating the latest deep learning models. The applied scientist optimizes different models for specific platforms, including edge devices with restricted resources. Generative AI, Multi-modal models, e.g., Large Vision Language Models, zero-shot, few-shot, and semi-supervised learning paradigms are used extensively.