Looking for a Technical Co-Founder — Large Action Model for Embodied AI / Robotics
I’m an independent researcher based in West Africa, currently developing a Large Action Model (LAM) designed to provide a general-purpose intelligence layer for robots.
The core idea is to move beyond the paradigm where an LLM directly generates executable robot actions.
My architecture is based on:
* Multimodal perception and world understanding
* Reasoning and action verification
* Skill acquisition through imitation and reinforcement learning
* A reusable library of learned robotic skills
* Dynamic composition of skills into executable workflows
* Continuous learning from environmental feedback
* Robot-agnostic action execution
The project is called Action-Workflow. The long-term objective is to build an intelligence layer that allows robots to acquire and reuse physical skills instead of solving every task from scratch.
I am looking for a technical co-founder with strong experience in one or more of:
* Robotics / embodied AI
* Reinforcement or imitation learning
* Multimodal models
* Robot learning and simulation
* LLM/agent architectures
* Computer vision / world models
* Distributed AI systems
I am particularly interested in someone who wants to work on the research and engineering problem itself, not simply take on a conventional software-development role.
I am open to a long-term remote collaboration and an equity-based co-founder arrangement, depending on the profile and level of commitment.
I can provide the current architecture, technical documentation, research direction and existing implementation privately for serious candidates.
I am French-speaking and based in West Africa, but the project is intended to be international.
If this problem interests you and you are seriously considering building a company around embodied intelligence, feel free to comment or DM me.