Impress employers and recruiters.
Choose from hundreds of resume examples.

Target Resume Now

About the Role

You will own the control interface of Luma's video foundation models, focusing on fine-tuning, adapters, and personalization to enable precise direction by creative partners. This role bridges research, product, and partnerships, translating model capabilities into production needs. You will work full-stack across modeling, data, systems, and evaluation, with an emphasis on personalization and long-horizon memory to make models understand specific workflows. This position is ideal for a researcher who values collaboration with partners and focuses on high-fidelity problems over benchmark scores.

Responsibilities

Provide precise control and creative range for Luma's models in high-fidelity, partner-grade workflows using SFT, RL, distillation, and adapter-based methods.
Embed domain expertise and long-horizon memory into models through context management, PEFT, and preference learning.
Ground the data engine in real-world workflows, transforming usage into training data and insights for future model development.
Define and drive end-user quality by setting success metrics, building user-aligned evaluations, and iterating the model/data/eval loop to strict fidelity targets.
Collaborate with Product and Design to translate creative intent and feedback into clear specifications and shippable model behaviors.

Requirements

Strong ML foundation with deep experience in visual generative models (diffusion, transformers, or related).
Depth in at least one of: fine-tuning, personalization, domain adaptation, data curation, targeted distillation, interpretability, or human-feedback refinement.
Hands-on experience with PyTorch and large model training.
A product instinct: treat users and partners as collaborators and solve their specific problems.
Contributions to state-of-the-art image or video generation models (Nice to Have).
Experience working with creative partners (VFX, animation, film, design tools) (Nice to Have).
A track record building workflows or tools that improve iteration speed and evaluation rigor (Nice to Have).
Familiarity with large-scale training infrastructure and distributed systems (Ray, Slurm, Kubernetes) (Nice to Have).

Skills

Fine-tuning
Personalization
Domain Adaptation
Data Curation
Targeted Distillation
Interpretability
Human-Feedback Refinement
PyTorch
Large Model Training
Visual Generative Models
Diffusion Models
Transformers
SFT
RL
Context Management
PEFT
Preference Learning
Evaluation Metrics
Distributed Systems
Ray
Slurm
Kubernetes

Work Type

Full-time

About the Company

Luma's mission is to build unified general intelligence that can generate, understand, and operate in the physical world.
We believe multimodality is critical for intelligence — the next step beyond language models comes from vision.

Equal Opportunity

Luma is an equal opportunity employer.