LLM Engineer - Agents

About Dimensional

Dimensional is building the foundation for physical intelligence in general-purpose robotics. We develop a unified framework that enables robots to perceive, reason, learn, and act across diverse real-world environments and embodiments. Dimensional’s mission is to make physical intelligence programmable, scalable, and accessible -- accelerating the development of generalist robots that can adapt, learn, and perform reliably in the real world.

Our platform sits between robot hardware and AI models, providing a modular, composable architecture for training and deploying robotic intelligence at scale. By abstracting hardware differences and standardizing core capabilities—such as perception, spatial understanding, decision-making, and control—Dimensional allows developers and partners to build, transfer, and operate robotic skills efficiently across platforms.

About The Role

We’re hiring an LLM Engineer to build intelligent agent systems that interact with real-world robotic platforms. This role sits at the intersection of cutting-edge agent development, multimodal reasoning, and production-grade software engineering.

You’ll help develop an advanced agent harness designed to rival leading coding agents like Claude Code and Codex, while leveraging the unique capabilities of Dimensional OS. Unlike traditional coding agents, our systems must reason over massive volumes of real-world sensor data, extract meaningful insights, and answer complex, context-dependent questions, all while efficiently managing context windows and maintaining performance across long-horizon tasks.

We're looking for engineers who thrive at the intersection of research and production, with the ability to experiment rapidly, solve complex technical challenges, and ship reliable systems, not just prototypes. You might be a former founder, an early engineer at an agent-focused startup, or someone deeply immersed in the modern agent ecosystem with exceptional software engineering fundamentals.

A few examples of what you might work on

• Evaluate agent performance by modeling real-world robotics scenarios in simulation environments such as MuJoCo and Habitat.

• Develop advanced agent harnesses capable of processing massive volumes of sensor data while maintaining high performance across long trajectories.

• Build multimodal reasoning and fact-finding systems that integrate point clouds, images, audio, and text to extract meaningful insights.

• Design agent memory systems that preserve relevant information across runs, including memory migrations, context organization, and orchestration state management.

• Develop holistic representations of 3D environments using scene reconstruction and spatial reasoning techniques, enabling agents to efficiently query, interpret, and reason about their surroundings.

Core Responsibilities

• Design, build, deploy, and maintain single-agent and multi-agent systems for real-world robotics applications.

• Develop multimodal agent architectures capable of processing complex sensor data, understanding spatial environments, and interacting with robotic platforms.

• Build and optimize agent harnesses, including context management, persistent memory systems, and agent orchestration.

• Develop scalable systems for scene reconstruction, spatial reasoning, and efficient retrieval of information from complex environments.

• Integrate agent systems into production robotics workflows, ensuring reliability, scalability, and performance.

• Contribute high-quality, production-ready code to a large-scale, shared engineering codebase.

What We're Looking For

• Hands-on experience designing, building, and deploying LLM-powered agent systems, including single-agent and multi-agent architectures.

• Strong software engineering fundamentals, with the ability to deliver reliable, scalable, production-quality code within complex, shared codebases.

• Experience deploying and maintaining agent systems used by real customers, with an understanding of the technical challenges, failure modes, and performance bottlenecks that emerge at scale.

• Experience with agent harness engineering, including agentic memory, context management, orchestration, and long-horizon reasoning.

• Ability to design rigorous evaluation frameworks, define quantitative success metrics, and systematically improve agent performance through experimentation and optimization.

• Familiarity with modern agent frameworks and tooling, combined with strong product intuition and an understanding of how users interact with intelligent agents.

Bonus (not required)

• Experience as an early-stage founder or engineer at an agent-focused or AI-native startup.

• Deep familiarity with contemporary agent research, frameworks, and emerging AI tooling.

• Proven experience translating experimental or research-driven AI systems into production environments.

• Experience working with robotics sensor data, including LiDAR, point clouds, joint states, spatial representations, and feature embeddings.

• Familiarity with scene reconstruction, 3D computer vision, or robotics simulation environments such as MuJoCo and Habitat.

• Experience with reinforcement learning, automated evaluation, or optimization techniques for improving agent performance.

Dimensional is an equal opportunity employer committed to diversity and inclusion, supporting all applicants without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, Veteran status, or other legally protected characteristics.

At Dimensional, accessibility is a fundamental human right. By welcoming diverse perspectives, we help you build a career where you feel you belong.