What are the responsibilities and job description for the Qualitative Evaluation Engineer position at Luma?
About Luma AI
Luma's mission is to build multimodal AI to expand human imagination and capabilities. We believe that multimodality is critical for intelligence. To go beyond language models and build more aware, capable, and useful systems, the next step function change will come from vision. So, we are working on training and scaling up multimodal foundation models for systems that can see and understand, show and explain, and eventually interact with our world to effect change.
About The Role
Luma is pushing the boundaries of generative AI, building tools that redefine how visual content is created. We’re seeking a candidate to help shape and scale the way we understand, measure, and improve model performance. In this role, you’ll partner with researchers, engineers, and technical artists to evaluate our models against real-world creative use cases, design frameworks that capture qualitative nuance, and identify actionable insights that guide development.
This is not a checkbox metrics role - it's about building evaluative systems that match the complexity of human perception, creativity, and intention.
Responsibilities
Luma's mission is to build multimodal AI to expand human imagination and capabilities. We believe that multimodality is critical for intelligence. To go beyond language models and build more aware, capable, and useful systems, the next step function change will come from vision. So, we are working on training and scaling up multimodal foundation models for systems that can see and understand, show and explain, and eventually interact with our world to effect change.
About The Role
Luma is pushing the boundaries of generative AI, building tools that redefine how visual content is created. We’re seeking a candidate to help shape and scale the way we understand, measure, and improve model performance. In this role, you’ll partner with researchers, engineers, and technical artists to evaluate our models against real-world creative use cases, design frameworks that capture qualitative nuance, and identify actionable insights that guide development.
This is not a checkbox metrics role - it's about building evaluative systems that match the complexity of human perception, creativity, and intention.
Responsibilities
- Evaluate generative model performance across diverse tasks, prompts, and modalities.
- Identify key failure modes, regression patterns, and edge cases that impact product quality.
- Develop and maintain qualitative evaluation frameworks that are scalable and reusable.
- Collaborate closely with technical artists and engineers to align evaluations with model capabilities and target use cases.
- Translate high-level product goals into concrete evaluative criteria.
- Lead qualitative studies, side-by-side comparisons, and human-in-the-loop evaluation efforts.
- Provide detailed feedback that informs model fine-tuning, dataset curation, and product UX.
- Stay informed about emerging evaluation standards in generative AI and creative tools.
- Master’s degree or higher in Cognitive Science, Human-Computer Interaction (HCI), Design Research, Psychology, Media Studies, or a related field.
- 5 years of experience in product evaluation, UX research, model testing, or similar roles that involve structured qualitative assessment.
- Deep familiarity with creative workflows and real-world use cases for generative models (e.g., animation, filmmaking, digital art, VFX).
- Strong systems thinking and the ability to define abstract qualities (like believability, identity retention, or scene coherence) in clear evaluative terms.
- Experience working cross-functionally with engineers, researchers, and creatives.
- Excellent written communication skills and the ability to synthesize nuanced judgments into clear, actionable insights.
- Background in motion, visual effects, or storytelling pipelines
- Experience evaluating AI-generated media (video, images, 3D)
- Prior work on building internal tools for qualitative data collection or scoring
- Familiarity with prompt engineering and reference-based input methods