Backend Engineer | Multimodal AI Systems

Luma AI is building the next era of AI with Omni models that can see, hear, and understand the world. As a full-stack company, we train our own foundational models and build the products that utilize them. We operate with the capital and compute resources necessary to compete at the frontier of AI while maintaining a lean team structure. You will build the intelligence layer that powers autonomous agentic workflows and massive-scale inference, designing systems for generative AI complexity, managing inference pipelines, and building infrastructure for autonomous agents. You will collaborate with the research team to productionize novel capabilities.

Must Have

• Build the backend systems that enable autonomous AI agents to perform complex, multi-step creative tasks.
• Design high-throughput systems capable of serving generative video and audio to millions of concurrent users, solving novel challenges in job queuing and media processing.
• Build the serving layer for our proprietary multimodal models, optimizing for inference speed and reliability.

The Opportunity

Luma AI is building the next era of AI with Omni models that can see, hear, and understand the world. As a full-stack company, we train our own foundational models and build the products that utilize them. We operate with the capital and compute resources necessary to compete at the frontier of AI while maintaining a lean team structure that guarantees you will be part of the core engineering group solving the hardest problems.

Where You Come In

You will build the intelligence layer that powers our autonomous agentic workflows and massive-scale inference. This role involves designing systems that can handle the extreme complexity of generative AI, from managing inference pipelines to building the infrastructure for autonomous agents. You will work directly with our research team to productionize novel capabilities.

What You Will Build

• Agentic Infrastructure: Build the backend systems that enable autonomous AI agents to perform complex, multi-step creative tasks.
• Scale and Reliability: Design high-throughput systems capable of serving generative video and audio to millions of concurrent users, solving novel challenges in job queuing and media processing.
• The Intelligence Layer: Build the serving layer for our proprietary multimodal models, optimizing for inference speed and reliability.

The Profile We Are Looking For

• Technical Judgment: You have a history of making high-stakes technical decisions for complex systems, demonstrating the engineering judgment required to balance speed, reliability, and scale in a production environment .
• Systems Thinker: You have a track record of building scalable, distributed systems from scratch. You prefer inventing solutions for novel problems over maintaining existing platforms.
• Research Collaboration: You are comfortable operating in a fast-paced environment where engineering influences research, and want to be in the room where core decisions are made.
• Technical Depth: Expert-level fluency in Python, with strong experience in Kubernetes, distributed systems, or AI frameworks.