Machine Learning Engineer, Stripe Assistant

Stripe Stripe · Fintech · United States · 8212 ML Foundations

Stripe is seeking a Senior Machine Learning Engineer to own the end-to-end ML and agent architecture for the Stripe Assistant. This role involves developing an intelligent assistant that leverages LLMs and agentic systems to answer queries, resolve issues, provide business insights, and anticipate user needs. Responsibilities include setting strategy for high-trust actions, delivering accurate answers, orchestrating tools and agents, grounding responses in data, driving conversation continuity, establishing evaluation and SLOs, and improving quality, latency, cost, and availability. The role also involves mentoring engineers and upholding high standards for code quality, security, and operational rigor.

What you'd actually do

  1. Establish trustworthy, human-in-the-loop execution for high-trust “write” actions—prioritizing user control, transparency, accountability, and auditability so customers can delegate with confidence.
  2. Define and evolve the Assistant’s capability and governance model across hundreds of tools and agents, balancing power, permissions, and consistency at scale.
  3. Raise answer quality and usefulness by grounding in authoritative Stripe knowledge and live user data, building cross-surface memory and personalization, and making the Assistant proactive and present in the dashboard.
  4. Explore and apply optimal machine learning methods to improve Stripe Assistant’s overall performance, including but not limited to fine-tuning LLMs with RLHF, synthetic data generation, optimizing RAG pipelines via domain‑specific embedding and retriever fine‑tuning, and automatic prompt tuning, etc.
  5. Make quality and reliability a product: set and meet SLOs, build rigorous evaluation and benchmarking loops, and drive sustained improvements in latency, cost, and availability.

Skills

Required

  • 5+ years in AI/ML and backend engineering
  • Applied LLM experience: RAG/embeddings, tool use/function calling, agentic planning/orchestration, fine-tuning, code generation, evaluations, etc.
  • Proficient in Python (Ruby is a plus)
  • strong distributed systems fundamentals
  • Experience working closely with product management, design, other engineers, and other cross-functional partners

Nice to have

  • Experience operating ML systems at global scale with stringent SLOs—balancing reliability, latency, and cost—with privacy, security, and compliance by design.
  • Experience building products where AI/ML is core; as well as balancing short-term product priorities with long-term AI/ML improvements.
  • Track record building ML platforms, especially those that enable multiple teams to collaborate together.
  • Strong technical leadership and communication: mentoring and elevating engineers, elevating AI/ML awareness and posture within organizations, setting architectural direction, and driving alignment in ambiguity.

What the JD emphasized

  • high-trust actions
  • rigorous evaluation
  • SLOs
  • quality, latency, cost, and availability
  • trustworthy
  • proactive AI operating layer
  • Applied LLM experience: RAG/embeddings, tool use/function calling, agentic planning/orchestration, fine-tuning, code generation, evaluations, etc.
  • stringent SLOs

Other signals

  • building an intelligent and proactive assistant that not only answers users’ queries but efficiently resolves issues and provides valuable business insights
  • leverage LLMs and agentic systems
  • evolve the Assistant into a proactive partner that anticipates user needs
  • establish rigorous evaluation and SLOs
  • deliver step‑change improvements in quality, latency, cost, and availability
  • configurable levels of autonomy
  • trustworthy, proactive AI operating layer for every Stripe merchant