Research Engineer — Reinforcement Learning

About this role

Firecrawl seeks a Research Engineer to build reinforcement learning systems that improve their web data extraction product. You'll design training infrastructure, reward pipelines, and fine-tuning systems while bridging classical RL with modern LLM agents, shipping models to production at a fast-moving startup.

What you'll do

• Design and operate training infrastructure, data pipelines, and reward modeling systems from scratch
• Fine-tune foundation models to achieve state-of-the-art performance on web data extraction tasks
• Apply reinforcement learning techniques to improve multi-step LLM agent workflows
• Run rapid experiments to validate hypotheses and iterate on model improvements
• Communicate RL research findings clearly to non-technical stakeholders and product teams
• Collaborate with IR-focused researchers and engineers to integrate RL improvements into the product

What they're looking for

• Reinforcement learning (PPO, RLHF, reward modeling, policy optimization)
• Model fine-tuning and training infrastructure
• GPU cluster management and training operations
• LLM agents and prompt engineering
• Data pipeline and evaluation framework design
• Production ML systems and deployment
• Fast experimentation and iteration
• Clear technical communication

Benefits

• Competitive salary: $180,000–$290,000/year plus up to 0.15% equity
• Flexible work: San Francisco hybrid or remote (Americas, UTC-3 to UTC-10)
• Small, technical, fast-moving team
• Opportunity to ship research directly to production
• Work on essential infrastructure for LLM applications