Senior Research Scientist | Model Steering

📋 Description

• Drive the development of translation models steerable by user context.
• Lead post-training for core translation models: SFT, distillation, and RL.
• Build reward and evaluator models for translation; address reward hacking.
• Push models to ingest multimodal content and context to raise translation quality.
• Own full lifecycle of model delivery: prototyping, training, evaluation, deployment at scale.
• Establish practices for evaluation, reproducibility, monitoring, and continuous improvement.

🎯 Requirements

• Proven experience making large models steerable and instruction-following.
• Hands-on expertise in LLM post-training (SFT, DPO), distillation, and RL.
• Strong data-centric instincts for synthetic and preference data pipelines.
• Experience designing evaluation and reward signals using automatic metrics and human-in-the-loop.
• Hands-on builder who trains models, runs experiments, and ships ML into production.
• Ownership of a substantial research direction with mentoring on a fast-moving team.

🎁 Benefits

• Diverse and internationally distributed team.
• Open communication and regular feedback.
• Hybrid work with twice-weekly office visits.
• Virtual Shares and ownership.
• Regular in-person team events.
• Monthly full-day hacking sessions.