Machine Learning Engineer, Coach Platform

Let’s face it: if you’re going to build the future of human transformation, you deserve more than a paycheck.

You deserve a personal BetterUp Coach, the most curious teammates on the planet, and a mission that pulls you out of bed every morning. At BetterUp, we pair world-class coaching with Generative AI to help millions live with clarity and purpose. Ready to engineer the models that make that magic feel personal?

What you’ll do

• Prototype, fine-tune, and evaluate LLM-powered features—from conversational insights to practice-moment nudges—using Python and Jupyter notebooks.

• Own the production lifecycle: data prep, experiment design, offline/online evaluation, and weekly pushes to prod alongside a tight squad of ML/Backend/Full Stack Engineers.

• Collaborate daily with Product, Coaching Science, and Learning Design to turn behavior-change research into delightful user moments.

• Contribute to technical design by shaping model architectures, retrieval strategies, and safety guards—while staying hands-on in code.

• Measure what matters: define success metrics, implement A/B tests, and iterate quickly on real coach feedback.

• Ship end-to-end improvements to products like FocusFrame, our AI copilot that helps coaches deliver higher-impact sessions in real time.

What you’ll bring

Core

• 3–5 years building and shipping ML systems in production, with at least 6–12 months hands-on with LLM fine-tuning, prompt engineering, or RAG pipelines.

• Fluency in Python for AI development and comfort working in Jupyter/IPython.

• Solid understanding of modern ML tooling (e.g., PyTorch, TensorFlow, or JAX) and containerized/cloud deployment basics.

• Clear, approachable communication with both technical peers and non-technical partners.

• Bias toward action, curiosity, and comfort navigating ambiguity in a fast-moving startup environment.

Bonus (nice-to-have, not required)

• Prior experience in coaching, learning-science, ed-tech, or behavior-change domains.

• Exposure to RLHF, preference tuning, or multimodal models.

Not sure you hit every bullet? We value growth mindsets—please apply anyway.