Autodesk AI Lab — AI/ML Developer Intern

Worked on Autodesk's AI and innovation team building agentic AI systems for Autodesk Assistant, with a focus on making advanced multi-agent capabilities practical and approachable inside existing creative workflows.

Following Autodesk's February 18, 2026 announcement of a $200 million strategic investment in World Labs, integrated World Labs spatial intelligence into Maya and prototyped an end-to-end Gaussian-splat workflow for generating, previewing, importing, and positioning AI-created 3D worlds.

Designed multi-agent workflows on the Claude Agent SDK, treating the system as a graph of specialized agents rather than a single prompt: each node owns a slice of the task, and the architecture centers on how context is managed and passed as agents hand work off to one another. This graph-engineering approach made it possible to reason about and debug individual agent responsibilities independently.

Built an in-Maya World Labs workflow that lets artists create 3D worlds from image or text inputs, preview generated panoramas, import Gaussian splats and collision meshes, and lock the Maya camera to the generated scene — keeping spatial-AI tooling inside the familiar Maya workspace to reduce setup and context switching.

Shipped the work as an Electron desktop application through a Jenkins CI/CD pipeline that built and signed macOS DMG installers for internal distribution, enabling frequent, low-friction testing cycles ahead of leadership demos.

Delivered functional, testable builds spanning Maya, Electron, multi-agent orchestration, and CI/CD. Demoed the agentic Assistant workflows and Maya integration directly to senior leadership as a high-visibility initiative within the AI and innovation team.

Operated in two-week scrum sprints with daily standups, working cross-functionally with principal engineers, senior architects, project managers, and senior UI/UX designers to move the project from prototype to a shippable, testable product.