AI Engineer / Machine Learning Engineer — 0.01% Equity

RoleColorFinder™ is a fast-growing leadership development and personality-mapping platform that helps schools, organizations, and individuals discover their natural leadership colors. We combine psychology, design, and data to build systems that develop adaptable, values-driven leaders for the future. The Role We’re looking for an AI / Machine Learning Engineer to build the intelligence layer behind the RoleColorFinder leadership engine. You’ll work on everything from prompt engineering and model orchestration to embeddings, vector search, and our AI-driven Slack assistant. Your work will directly shape how users experience personalized insights, coaching, and leadership recommendations. This is an unpaid, part-time, equity-only role at this stage. This role is perfect for someone who loves rapid experimentation, working across multiple layers of the AI stack, and building features that ship fast. Responsibilities Develop prompt pipelines, agents, and lightweight models Build personalized leadership insight generation tools Implement embeddings and vector search within Supabase Power the AI-driven Slack assistant and reasoning workflows Create internal tools for automated summaries and analytics Iterate quickly, test outputs, and optimize for reliability and clarity Requirements Solid understanding of LLMs, prompt engineering, and OpenAI API usage Experience with Python or JavaScript Bonus: LangChain, embeddings, Supabase vector search, agents, tool-calling workflows Ability to design systems that are reliable, testable, and predictable Comfort working in a fast-paced, early-stage environment ⏰ Commitment Flexible hours (part-time expectations) Fully remote, async-friendly High ownership over AI systems and experimentation Compensation 0.01% equity Performance-based scale-up to 0.05% Note: Although Wellfound requires this role to be listed as “Full-Time,” the actual workload is flexible and part-time in practice. We care about consistent communication and strong output, not strict scheduling. Apply tot his job