SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
LM Studio is building tools that enable millions of people to run AI on their own computers, with Bionic extending capabilities to the cloud. The company prioritizes human-centered design and creating products the team would personally recommend.
As a Software Engineer on the Agent Harness team, you will build the core agentic system powering Bionic, LM Studio's agent for work and code. Your focus will be making open models effective at working with codebases, documents, files, software tools, and the user's computer.
Open models vary significantly in reasoning, tool calling, context usage, and error recovery. You'll develop the software infrastructure that handles these differences and extracts optimal agentic performance from models running locally, on linked devices, and in LM Studio's inference cloud.
Key responsibilities include building and improving the agent harness powering Bionic; making local and cloud models reliable at using files, search, Git, shell, web tools, and document tools; integrating new open models and optimizing their performance; managing conversation history, working context, tool results, interruptions, and recovery; building repeatable evaluations and diagnostics for failure analysis; constraining models intelligently to maintain usefulness; and collaborating across Bionic, the LM Studio runtime, and cloud infrastructure.
You should bring significant production experience with language models, developer tools, or automation systems. Strong understanding of tool calling, structured model output, context construction, and streaming inference is essential. Experience building stateful systems that perform multi-step work and recover cleanly from failures is required. Excellent TypeScript and general software engineering fundamentals are expected. You need deep interest in open models and practical understanding of their strengths and limitations, combined with high technical intensity and personal responsibility for finished products.
