AI Operator
Predict.fun · Remote · Mid
About the role
Predict.fun is looking for an AI Operator to build LLM-driven tools and workflows. You will focus on building robust control flows where LLMs return structured outputs that directly drive application logic, orchestration, and multi-step reasoning. This is a technical role that blends software engineering and LLM prompting. You will help create systems where AI agents and models help automate or speed up complex tasks. What You'll Do: - Design and implement AI workflows using structured outputs — Provided some requirements and guidance, you will be responsible for building and maintaining tools and workflows, either standalone or integrated with our backend systems, that help automate complex tasks. - Collaborate with the team to improve reliability, quality, and cost-efficiency of LLM-powered features — The tools and workflows you contribute to will need monitoring to make sure they stay useful, reliable and cost-effective. Both expense monitoring and tracking key metrics around error rates will be an important part of this role.
Requirements
- AI Workflow Experience: Practical experience coding AI/LLM applications, particularly using structured outputs to control application logic and multi-step workflows (e.g., via OpenRouter, OpenAI, Anthropic, Groq, or similar APIs) - 1–2+ years of software engineering experience, preferably backend - Solid understanding of PostgreSQL - Proficiency with Rust and/or Node.js (TypeScript); Rust is preferred, other lower-level language experience such as C++ or Go can substitute for Rust - Blockchain Understanding: Working knowledge of blockchain concepts (smart contracts, wallets, on-chain transactions, etc.) Nice to Have: - Experience or familiarity with prediction markets (e.g., Polymarket, Kalshi, or similar platforms)