AI Research Manager | Translation Quality

DeepL DeepL · AI Frontier · Munich, Germany · Research

AI Research Manager to lead the Translation Quality Research team, focusing on advancing and defending core translation models. The role involves shaping research agendas, prioritizing research, defining strategic goals, building models for human review estimation, and leading/developing a team of research scientists. The emphasis is on measurable quality and real-world production impact to maintain market dominance against large general-purpose models.

What you'd actually do

  1. Drive the achievement and rigorous evaluation of DeepL's baseline translation quality to ensure market dominance.
  2. Shape and prioritize the research agenda, balancing clear product outcomes with bold, longer-term bets.
  3. Define and communicate your team's strategic goals, roadmap, and success metrics in collaboration with key stakeholders across tech, product, sales enablement, and customer success.
  4. Own the "estimation" aspect of our product pipeline, building models that determine when a translation requires human review.
  5. Inspire, develop, recruit, and motivate a high-performing team ; build strong 1:1 relationships, set development plans, and foster a candid, non-retaliatory feedback culture.

Skills

Required

  • Technical and people leadership
  • Deep Learning
  • Software development, ideally in the context of data science, ML engineering, systems engineering, or distributed services
  • Strong research background—ideally with specific experience in Language Research
  • Excellent communication skills
  • Solution-oriented mindset
  • Ability to navigate through ambiguity and change effectively
  • High level of English proficiency

Nice to have

  • PhD degree in Computer Science, Mathematics, Physics, or a comparable subject
  • PhD focused on a heavy research topic
  • Experience in a global, high-growth scale-up, and research-intensive environment
  • Knowledge of German and other languages

What the JD emphasized

  • rigorous evaluation
  • measurable quality
  • real-world production impact
  • market dominance
  • highest standard of translation quality
  • estimation
  • building models that determine when a translation requires human review
  • high-performing team
  • deeply understand complex tech challenges
  • grasp the team's constraints
  • proven track record of successfully leading Research teams

Other signals

  • advancing core translation models
  • keeping baseline translation quality ahead of massive, top-tier general-purpose models
  • driving strategic enterprise initiatives
  • delivering the highest standard of translation quality to B2B customers
  • building models that determine when a translation requires human review