TL;DR

Product Manager (AI): Driving the research roadmap for Decagon’s AI agent development, focusing on model training, evaluation systems, and agent architecture. Focus on marrying cutting-edge technology with key customer problems and ensuring Decagon’s AI capabilities lead the market.

Location: On-site in San Francisco

Salary: $200,000 – $285,000 + Equity

Company

Decagon is a leading conversational AI platform empowering brands to deliver concierge customer experiences, backed by world-class investors.

What you will do

  • Define the strategy and roadmap for the Research team, including model training, evaluation, and experimentation.
  • Work closely with customers to deeply understand their problems and requirements.
  • Translate customer problems into clear model requirements and partner with engineering to implement them.
  • Interface tightly with Engineering, Infrastructure, Agent PM, and Go-to-Market teams.
  • Manage relationships with key external partners.

Requirements

  • 5+ years of relevant experience (e.g., product manager, founder, engineer, researcher).
  • Experience developing AI/ML models, particularly in a research environment.
  • Strong technical acumen in model architecture, training data, and evaluation pipelines.
  • Strong communication and relationship-building skills.
  • Proven track record of shipping impactful features in fast-paced environments.
  • Must work on-site in San Francisco.

Nice to have

  • A Computer Science, Engineering, or Math degree or equivalent technical experience.
  • Experience working at a startup or high-growth company.
  • Experience working on an enterprise B2B product.

Culture & Benefits

  • Comprehensive medical, dental, and vision benefits.
  • Flexible "take what you need" vacation policy.
  • Daily lunches, dinners, and snacks provided in the office.
  • Work in an in-office company driven by a shared commitment to excellence and velocity.
  • Values include "Just Get It Done", "Invent What Customers Want", "Winner’s Mindset", and "The Polymath Principle".