Founding Machine Learning Engineer at Withshepherd

As the first Machine Learning Engineer, you will build the foundation of a fully autonomous underwriting system for high-hazard industries. You will design and ship production ML systems that transform static construction data into real-time risk assessments, moving the industry toward an agentic future where complex insurance submissions are priced in seconds without human intervention.

• Design, build, and ship production-grade ML systems and agentic LLM workflows that power autonomous underwriting decisions.
• Build and close the feedback loops that translate human underwriter expertise into training signals and compounding model improvements.
• Develop rigorous confidence scoring and evaluation frameworks to determine when the system can take on more autonomy versus needing human review.

• 4+ years of industry experience building end-to-end ML systems, from raw data processing to production deployment via platforms like AWS SageMaker.
• Deep technical proficiency in Python and Pytorch with specific experience fine-tuning SLMs/LLMs using techniques like RLHF, DPO, or LoRA.
• Proven track record of shipping LLMs in production, including prompt engineering, tool use, and building reliable models with limited labeled data.

• Lead the charge toward the first fully agentic submission in the industry, essentially building the "Waymo for underwriting" for physical world infrastructure.
• Backed by over $60M in funding, including a recent $42M Series B led by Intact Private Capital, providing massive capital and industry-leading carrier partnership.
• Massive ownership as the first ML hire, with the authority to define the ML lifecycle, platform, and registry from the ground up at a high-growth startup.