Founding ML Engineer

Location: San Francisco, CA Work Model: In-office 5 days a weekAbout UsAt Effective AI, we're building the future of work. We believe the real frontier for AI is not simple, repetitive tasks, but complex knowledge work that demands expertise and multi-step reasoning. That's why we're building sophisticated AI Teammates designed to master complex workflows and collaborate with human experts. We're starting by tackling the trillion-dollar U.S. Property & Casualty insurance market, a space with intricate processes and a rich data environment perfect for this challenge.We’ve raised $10 million in seed funding from Lightspeed Ventures & Valor Equity Partners.Our dedicated team is based in San Francisco and thrives on in-person collaboration to solve these challenging problems.What You'll DoAs a Founding ML Engineer, you will be a crucial part of our initial team, playing a pivotal role in designing, post-training, and deploying the agent loops that power our AI Teammates from the ground up. You will tackle some of the most significant challenges in agentic AI and natural language processing to develop AI Teammates capable of handling core insurance functions like underwriting and claims processing.More specifically, you will:
Architect and Develop Core ML pipelines: Design, train, and fine-tune state-of-the-art language models (including reinforcement learning agents) to enable long-horizon task completion and complex decision-making.
Implement Nuanced Reasoning: Develop and integrate ML techniques that empower agents to make sound judgments on ambiguous or incomplete data, mimicking human expert reasoning and generalization.
Build Intelligent, Tool-Using Agents: Engineer the ML systems that allow our agents to dynamically select and effectively utilize a diverse set of external tools—including APIs, databases, web searches, and even Excel-based pricing algorithms—to gather facts and take action.
Design and Implement Robust Evaluation Frameworks: Develop and utilize comprehensive evaluation metrics and systems to rigorously test and benchmark agent performance, identify areas for improvement, and ensure reliability and safety in real-world insurance workflows.
Design for Continuous Adaptation and Learning: Create robust ML pipelines and feedback loops that enable agents to learn from new data, adapt to dynamic conditions like changing regulations or market shifts, and continuously improve performance.
Establish MLOps Best Practices: Contribute to the foundational infrastructure for model development, deployment, monitoring, and iteration in production environments.