Open Role
Research Engineer: Robotics Agent
Singapore / Munich, Germany
Build robotics agents that understand tasks, plan actions, invoke skills, and adapt their behavior based on execution feedback.
About the Role
MindOn is building general-purpose intelligence for robots. We are looking for a Research Engineer to design, develop, and deploy robotics agents on real systems.
You will combine large language models (LLMs), vision-language models (VLMs), and robot skills to translate user intent into executable tasks, understand changes in the environment, recover from failures, and work toward completing goals.
Responsibilities
- Design and develop robotics agent architectures that connect task understanding, environment perception, task planning, skill invocation, and execution feedback.
- Use LLMs and VLMs for task decomposition and multi-step planning, translating natural language instructions into executable robot workflows.
- Design tool-calling and skill interfaces that connect perception, navigation, manipulation, and other system capabilities.
- Build execution state management and feedback mechanisms for progress tracking, failure detection, retries, and replanning.
- Develop context and memory management methods that organize environment information, interaction history, and execution results to support long-horizon tasks.
- Collaborate with robotics algorithm, control, and software teams to integrate and deploy agents on real robot systems.
- Build task-level evaluation workflows and analyze success rates, planning quality, response latency, and failure causes to continuously improve performance.
Qualifications
- A deep and genuine passion for Embodied AI is essential.
- Proficiency with AI coding tools is essential: use them effectively for development, debugging, testing, and troubleshooting; independently review generated code and validate its correctness and reliability through tests and real-world execution.
- A bachelor’s degree or above in Computer Science, Robotics, AI, Automation, or a related field, or equivalent practical experience.
- Familiarity with the principles, capabilities, limitations, and use of mainstream LLMs and VLMs, with experience integrating, deploying, or developing applications with these models.
- Understanding of core agent mechanisms, including tool calling, structured outputs, task planning, context management, and memory, with practical system development experience.
- Ability to design reliable tool interfaces and execution workflows that handle uncertain model outputs, failed calls, timeouts, and inconsistent state.
- Understanding of the relationship between robot task planning and low-level skill execution, with the ability to assess plan feasibility against environment state and skill constraints.
- Proficiency in Python and strong software engineering skills, including asynchronous tasks, state management, logging, and testing.
- Rigorous experimental and evaluation skills, using execution traces and failure cases to diagnose issues and validate improvements.
- Strong problem-solving skills, independent execution, and cross-functional collaboration.
Preferred Experience
- Experience developing robotics agents, embodied AI systems, or integrating real robot systems.
- Experience with visually grounded planning, execution monitoring, or failure recovery.
- Familiarity with behavior trees, state machines, Task and Motion Planning (TAMP), or related methods.
- Familiarity with ROS/ROS2 and interface design for perception, navigation, or manipulation skills.
- Experience fine-tuning, deploying, or optimizing inference for LLMs or VLMs.
- Experience developing systems for long-horizon tasks, interactive task clarification, or human-robot collaboration.
- Relevant open-source projects or research contributions.