Computer Vision Annotation Product Strategy

Product challenge

Improve the economics and usability of complex computer-vision annotation. Customers needed precise training data, operators needed faster workflows, and internal project teams needed better visibility into turnaround time and feature adoption.

Product and engineering approach

• Researched the needs of annotators, project teams, and customers, then translated them into product requirements for 2D and 3D video workflows.
• Used my prior hands-on engineering experience with the annotation platform to assess feasibility, clarify trade-offs, and work effectively with the delivery team.
• Helped design a 3D segmentation tool that was presented at CVPR 2019.
• Improved internal project-management tools so operational teams could coordinate complex annotation work more effectively.
• Created dashboards for feature use, adoption, and turnaround time, giving roadmap decisions a stronger behavioral signal.

Outcome

• Reached more than 90% adoption for key features.
• Reduced annotation turnaround time by removing workflow friction identified through research and usage data.
• Established a more evidence-led product process spanning discovery, requirements, adoption, and iteration.
• Trained Product Solution Engineers on the new tools and relevant coding skills, helping product knowledge travel beyond the core team.