Senior Product Engineer, AI

Intercom Intercom · Enterprise · Berlin, Germany · AI Group

Senior Product Engineer role focused on building and shipping Intercom's AI-powered customer service products, working closely with ML Scientists and product teams to bring ML features from prototype to production. The role emphasizes product focus, rapid iteration, and measuring customer impact.

What you'd actually do

  1. Build the systems that power Intercom’s flagship AI products. This usually means working in our ML Python tech stack, but at times also means working across the product on our Rails app too.
  2. Work alongside our ML Scientists bring proof-of-concept code to production; ensuring it’s robust and scalable for real-world use.
  3. Partner with product teams outside of our group, to shape the best AI-powered product experience for our customers..
  4. Contribute to all phases of software development including ideation, prototyping, and implementation and testing.
  5. Play an active role in the hiring, mentoring, and career development of other engineers.

Skills

Required

  • 5+ years of experience shipping high-quality products
  • Deep knowledge of a high-level programming language (e.g., Ruby, Python, Javascript)
  • Strong willingness to fight for good outcomes
  • Bias towards progress over perfection
  • BSc in Computer Science, Maths or Statistics or related discipline

Nice to have

  • Previous experience in data analytics, or statistical role

What the JD emphasized

  • No ML experience required. We believe you can be highly effective immediately by bringing excellent software engineering skills, and learn ML as you work on the team.

Other signals

  • building AI-powered products
  • bringing proof-of-concept code to production
  • shipping to beta in weeks after a successful offline test
  • productized everything from classic supervised models, to cutting-edge unsupervised clustering algorithms, to novel applications of transformer neural networks
  • test and measure the real customer impact of each model we deploy