→ Luma's Applied Research team

→ First qual engineer at the company

→ Set the qualitative bar for multimodal generative AI quality across the company

→ Architect the evaluation frameworks that guide model training, product decisions, and launch readiness

→ Own and scale the human-evaluation pipeline as a core product and research capability

→ Drive data strategy by identifying systemic gaps in training and evaluation datasets

→ Serve as the internal authority on human behavior, emotion, and expression in generative outputs

→ Influence product, research, design, and GTM decisions

→ Product Strategy: Drove avatar direction through behavioral research and facial science; prioritized high-impact features and improved user satisfaction across Meta platforms, including Quest, Instagram, and AI Studio.

→ Research Leadership: Led 3+ simultaneous research initiatives at a time supporting product validation and system-level refinement across emerging, human-centered technologies.

→ Testing & Optimization: Executed iterative testing cycles to inform product decisions, improve system responsiveness, and advance real-time expression fidelity.

→ Evaluation Standards: Built and standardized FACS and UX-based frameworks used to evaluate expression fidelity and model alignment across 7+ high-visibility projects.

→ Cross-Functional Leadership: Aligned engineering, product, design, and cross-org teams by translating insights into roadmap shifts, product refinements, and system-wide cohesion.

Projects include:

→ Audio-to-Expression - model optimization, data integrity, QA authority

→ Codec Avatars - model evaluation, data collection

→ Horizon Avatars - avatar design, UX frameworks, model evaluation

→ AI Studio/ AI Agents - quality standards, model evaluation

Senior RA for Oculus VR - Face Tracking & Avatar Research