What I shipped

• Hook input normalization layer (cross-skill DX) for Rosetta
• AI-generated-file governance hook — catches stray files created by AI agents outside owning modules
• Sub-agent validation for independent code review (Rosetta Self-Automation epic)
• Front-end for Physical AI / robotics platform — 3D (Three.js) + visual workflows (xyflow/reactflow)

Challenge

Integrating AI into enterprise engineering workflows for a cutting-edge physical AI platform. The work spanned 3D browser interfaces (WebGL/Three.js) and robotics tooling — two technically distinct domains requiring simultaneous delivery under enterprise constraints.

Approach

AI-assisted development across both projects simultaneously, combining deep frontend expertise with AI-powered productivity:

• 3D visualization — WebGL/Three.js interfaces for robotics platform with complex spatial interactions
• AI backend integration — connecting frontend systems with AI/ML services for intelligent automation
• AI-powered tooling — leveraging AI coding assistants for development velocity, code review, and test generation
• Parallel delivery — managing 2 production projects with different tech stacks and release cycles

Results

• Delivered across two parallel production projects with different tech stacks and release cycles
• Transitioned to AI-first workflow by end of engagement
• Promoted to Staff Software Engineer based on delivery and technical leadership
• AI adoption spread across engineering workflows: coding, review, testing, documentation
• Demonstrated that enterprise AI adoption requires both tooling and process change

Key Insight