Large Language Model Systems Evaluation Engineer
Active AI foundation-model company offering multimodal models, consumer AI products, and an enterprise/developer API platform.
Funding history
About MiniMax Group Inc.
MiniMax develops proprietary multimodal foundation models spanning language, video, speech, and music, and operates AI-native products including MiniMax Agent, MiniMax Design, MiniMax Audio, Talkie, Hailuo AI, and an Open Platform.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will build evaluation scenarios based on real business and research problems. You will maintain environments, data, tools, MCP interfaces, harnesses, evaluators, and model integrations; run orchestrated and isolated evaluations; record execution artifacts; and support model comparisons, replay, regression detection, and troubleshooting.
Requirements
- Strong programming fundamentals and proficiency in Python.
- Experience in platform engineering, workflow orchestration, automated testing, or large-scale task execution.
- Ability to independently develop and maintain production systems.
- Familiarity with Linux, containers, networking, and distributed systems.
- Understanding of task scheduling, resource isolation, failure recovery, and observability.
- Ability to diagnose problems using logs, metrics, execution traces, and error samples.
- Basic understanding of large language models, agents, tool calling, and model evaluation.
- Engineering-quality awareness, documentation ability, and cross-functional collaboration ability.
Responsibilities
- Build evaluation scenarios that reproduce task flows, execution conditions, inputs, outputs, and completion criteria.
- Implement and maintain evaluation environments, data, tools, MCP interfaces, harnesses, evaluators, and model integrations.
- Manage versions and dependencies.
- Build evaluation execution capabilities for orchestration, sandbox isolation, resource scheduling, timeout control, retries, and batch runs.
- Record model requests, parameters, execution traces, tool calls, logs, metrics, rewards, and evaluation artifacts.
- Support replay and reproduction.
- Build model comparison and leaderboard management capabilities.
- Establish stability and maintenance mechanisms for regression detection, change-impact tracking, and issue diagnosis.