Applied Super Intelligence Engineer, Physical Systems

LLNL•Livermore, CA

1d•$121,830 - $185,544•Hybrid

About The Position

Lawrence Livermore National Laboratory is seeking an Applied Super Intelligence (SI) Engineer, Physical Systems, to build software that enables SI agents and multimodal models to interact with real-world equipment through an existing hardware integration stack. The role focuses on applied SI and perception workflows that transform operational data and equipment interfaces into safe, usable SI-assisted systems. It is in the Material Engineering Division (MED) within the Engineering Directorate and offers a hybrid schedule, with the flexibility to work from home one or more days per week. The role may be filled at either of two levels based on the hiring team's assessment of knowledge and related experience; additional responsibilities apply at the higher level. The engineer will build agent-facing tool definitions and workflows using documented equipment interfaces and platform services; develop perception and decision-support workflows using camera, telemetry, and process data; and improve prompt logic, tool use, evaluation, and workflow reliability. The position translates operational tasks into practical, testable, maintainable SI-assisted workflows and works with embedded, platform, and deployment engineers in a controlled environment to establish safe action boundaries, approval points, and fallback behaviors using established external safety and security systems. Workflows are evaluated and debugged through comprehensive logging, iterative refinement, and disciplined engineering practices. At the SES.2 level, the engineer also designs and evaluates workflows combining perception, tool use, and equipment actions across system components, modalities, and operational steps; improves reusable evaluation harnesses, guardrails, and failure analysis; and coordinates with embedded and SI platform engineers to turn device capabilities, requirements, and evaluation results into reusable workflows, while providing informal guidance to less-experienced contributors. This career-indefinite position requires the ability to obtain and maintain a DOE Q-level security clearance, which requires U.S. citizenship. The position may require work in Limited Areas where personal and laboratory mobile or wireless devices are restricted. External applicants selected must pass a post-offer, pre-employment drug test. The stated annual pay range is $121,830–$185,544, with separate ranges of $121,830–$154,500 for SES.1 and $146,340–$185,544 for SES.3.

Requirements

Ability to secure and maintain a U.S. DOE Q-level security clearance; U.S. citizenship is required for the clearance.
Bachelor’s degree in computer science, software engineering, robotics, or a related field, or an equivalent combination of education and related experience.
Proficiency in Python or a comparable programming language for building applications, prototypes, or internal tools in collaboration with software and systems engineers.
Experience gained through coursework, projects, research, internships, or professional work developing LLM, multimodal, agent-based, or tool-calling applications.
Experience designing, implementing, testing, and debugging tool-using SI workflows that connect models to APIs, services, data sources, or operational systems.
Experience with structured tool or API contracts, such as JSON Schema or OpenAPI, error handling, logging, and evaluation.
Experience working with real-world data from sensors, cameras, devices, or operational systems.
Familiarity with software development practices including version control, automated testing, and collaborative code review.
At the SES.2 level, ability to independently develop and evaluate multicomponent agent, multimodal, or tool-calling workflows for operational use.
At the SES.2 level, ability to scope and debug moderately complex workflow failures involving perception, tool use, system integration, and operational constraints.

Nice To Haves

Exposure to robotics workflows, ROS, or ROS 2.
Experience with vision-language models, vision-language-action models, or multimodal perception pipelines.
Familiarity with simulation, digital twins, or hardware-in-the-loop testing.
Familiarity with workflow orchestration, agent tracing or evaluation systems, prompt or configuration versioning, replay-based testing, or event and telemetry systems such as MQTT or NATS.

Responsibilities

Implement agent-facing tool definitions and task workflows using documented equipment interfaces and platform services.
Develop multimodal perception and decision-support workflows using camera, telemetry, and process data.
Improve prompt logic, tool use, evaluation methods, and reliability for physical-world SI applications.
Translate real-world operational tasks into practical, usable, testable, and maintainable SI-assisted workflows.
Collaborate with embedded, platform, and deployment engineers to define safe action boundaries, approval points, and fallback behaviors using established external safety and security systems.
Evaluate and debug applied SI workflows through comprehensive logging, iterative refinement, and disciplined engineering practices.
Perform other duties as assigned.
At the SES.2 level, design and evaluate agent workflows combining perception, tool use, and equipment actions across system components, input modalities, and operational steps.
At the SES.2 level, improve reusable evaluation harnesses, guardrails, and failure analysis, and use results to guide technical decisions.
At the SES.2 level, coordinate with embedded and SI platform engineers to develop reusable agent workflows from device capabilities, operational requirements, and evaluation results.
At the SES.2 level, provide informal guidance to less-experienced contributors.