Applied Researcher 3 - Staff

In the role of Applied Researcher 3 within the Global Focus Categories on the Applied Science team, you will develop new computer vision and multi-input AI solutions. These solutions will address complex problems across eBay’s global marketplace.

Our team focuses on challenges where current models and off-the-shelf methods fall short. We adapt and extend groundbreaking foundation models to solve specialized, domain-specific problems at scale.

The team works at the intersection of applied research, image and multimodal intelligence, and large-scale machine learning systems. You will work across a continually evolving portfolio of projects and areas of challenge rather than specializing in a single application area. Work has spanned personalized experiences and recommendations, content intelligence and attribute extraction, fine-grained visual recognition, anomaly detection, and multimodal understanding of marketplace inventory. Each problem brings different data, modeling, and system challenges, creating opportunities to explore new approaches and develop capabilities where established solutions fall short.

As a Staff level Applied Researcher, you will take problems from early investigation to production. You will research, build, train, fine-tune, evaluate, and deploy models and AI systems at scale. Collaborate with advanced image and multimodal core models. Develop specialized architectures, training approaches, data strategies, evaluation methods, and supporting systems to adapt them for challenging real-world applications.

As the Staff Applied Researcher on the team, your role includes helping develop the technical direction and identifying new opportunities for applied research. You will lead ambiguous problems from initial investigation through production impact, influence senior technical and product collaborators, and mentor researchers across the team.

Develop and deploy visual and multimodal machine learning models for large-scale applications involving fine-grained recognition, visual understanding, content intelligence, retrieval, and decisioning

Design and optimize end-to-end training and inference pipelines for modern AI systems, including orchestration of multiple models and services in production environments

Advance core capabilities in classical and modern vision and multimodal problem spaces such as detection, segmentation, classification, metric and contrastive learning, and visual-language modelin

Lead research initiatives focused on visual perception technologies and multimodal artificial intelligence. Identify gaps in which existing models are insufficient

Develop innovative strategies and adapt foundation models for specific domain problems

Lead research and development in VLM training and fine-tuning, model distillation, reinforcement learning, and other components of the modern AI stack

Partner closely with product managers, engineers, designers, and applied scientists to translate ambiguous business opportunities into robust technical roadmaps and delivered solutions

Communicate technical direction, trade-offs, and results clearly to cross-functional stakeholders, including director-level leaders, to drive alignment and informed decision-making

Mentor junior and senior team members, raise the technical bar for experimentation and production excellence, and contribute to a strong research and engineering culture