Senior Research Scientist | Multimodal Systems

DeepL
London, Cologne, Munich
Workplace: HybridFull timeFunction: Data Science & Machine LearningSkills: ["Communication","Collaboration","Experimentation","Debugging","Hands-on ownership"]

Lead fine-tuning, post-training, and reinforcement learning for the next generation of multimodal and vision models powering document translation. Develop models that reason about document layout using real-world and synthetic data, and make translation highly steerable and adaptable. Own the full ML lifecycle—prototyping through training, evaluation, optimization, and production deployment—while building evaluator models, improving reproducibility, and partnering with engineering to ship at scale.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
DeepL
DeepL
1 month ago

Senior Research Scientist | Multimodal Systems

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 8 hours agoStatus: Live

Job Summary

Lead fine-tuning, post-training, and reinforcement learning for the next generation of multimodal and vision models powering document translation. Develop models that reason about document layout using real-world and synthetic data, and make translation highly steerable and adaptable. Own the full ML lifecycle—prototyping through training, evaluation, optimization, and production deployment—while building evaluator models, improving reproducibility, and partnering with engineering to ship at scale.
Location: London, Cologne, Munich
Workplace: Hybrid
Employment Type: Full time
Job Function: Data Science & Machine Learning
Seniority: Mid level

Key Responsibilities

  • •Drive development of vision and multimodal models for document, image, and media translation from ingestion/generation to end-to-end models.
  • •Own hands-on post-training R&D including supervised fine-tuning, knowledge distillation, preference optimization, and reinforcement learning tuned to translation quality.
  • •Build evaluator models for document and design quality (rubric- and reference-based) and mitigate reward hacking and quality-estimation failure modes.
  • •Own the full model delivery lifecycle—prototyping, ablations, training, evaluation, optimization, and production deployment—partnering with engineering for real-time systems at scale.
  • •Establish strong practices for evaluation, reproducibility, monitoring, and continuous model improvement in production.

Key Requirements

  • •Proven experience developing multimodal models, VLM, and/or vision models.
  • •Deep hands-on expertise in model post-training, knowledge distillation, and/or reinforcement learning (RLHF/RLAIF, PPO/GSPO, reward modeling).
  • •Strong data-centric skills building synthetic-data and preference-data pipelines, model-as-judge generation, and data curation, filtering, and ablations.
  • •Hands-on experience training models, running experiments, debugging pipelines, and integrating ML systems into production with real-world quality focus.
  • •Strong coding and experimentation skills (Python, PyTorch/JAX/TensorFlow) and the ability to communicate and align research with product and engineering priorities.
Experience:Multimodal AIComputer visionReinforcement learningLanguage AIMachine translation
Skills:CommunicationCollaborationExperimentationDebuggingHands-on ownership
Tech Stack:PythonPyTorchJAXTensorflowRLHFRLAIFPPOGSPO

Company Brief

DeepL
DeepL builds Language AI products (DeepL Translator, DeepL Write, APIs and enterprise solutions) that provide high-accuracy translations and writing assistance to businesses and individuals, focusing on privacy, security and enterprise deployment.
Industry: AI & Machine Learning
Company Size: Enterprise (1,001+ employees)
Growth: Scaleup
Valuation: Unicorn (USD 1B+)
Funding: Series D
Headquarters: Cologne, Germany
Founded: 2017
Glassdoor
Glassdoor: 3.5
WebsiteLinkedInGlassdoor