Research Scientist - Post-Training Scope (Remote)

About the Role

We are seeking a Research Scientist - Post-Training Scope to join our elite team at Sentiro Partners. This is a unique opportunity to work in a world-leading research environment focused on building and aligning large-scale models under real constraints. As a Research Scientist - Post-Training, you will own the entire pipeline, directly observing and addressing failure modes at scale. This role is not just about academic theory; it's about applying your expertise in a practical, impactful way.

What You'll Do

• Lead research on how large models learn from human feedback and how reward signals shape behavior at scale.
• Design and optimize post-training infrastructure and RLHF (Reinforcement Learning from Human Feedback) loop.
• Investigate the dynamics of training and optimization behaviors in large language models.
• Develop robust RLHF pipelines that are resilient to distribution shifts.
• Collaborate with a small, elite team of researchers, sharing insights and driving innovation.

Requirements

• Post-graduate degree in machine learning, computer science, NLP, or a related field.
• 3-8 years of experience in post-training, alignment, or reward modeling.
• Published work in top conferences such as NeurIPS, ICML, ICLR, or ACL.
• Strong problem-solving skills and a deep curiosity about machine learning.
• Ability to work collaboratively in a high-performance research team.

Nice to Have

• Experience with preference optimization and training dynamics in large models.
• Familiarity with systems-aware ML where infrastructure decisions impact modeling choices.

What We Offer

• Competitive compensation and resources comparable to top-tier research labs.
• Unlimited compute resources and world-class reward packages.
• Deep technical autonomy and the opportunity to influence model training and alignment.
• Full visa sponsorship for international talent.
• Collaborative in-office environment on the East Coast.