Machine Learning Scientist (Post Training)

Machine Learning Scientist (Post Training)

Inflection AI is a public benefit corporation leveraging our world class large language model to build the first AI platform focused on the needs of the enterprise.

Who we are:

Inflection AI was re-founded in March of 2024 and our leadership team has assembled a team of kind, innovative, and collaborative individuals focused on building enterprise AI solutions. We are an organization passionate about what we are building, enjoy working together and strive to hire people with diverse backgrounds and experience.

Our first product, Pi, provides an empathetic and conversational chatbot. Pi is a public instance of building from our 350B+ frontier model with our sophisticated fine-tuning (10M+ examples), inference, and orchestration platform. We are now focusing on building new systems that directly support the needs of enterprise customers using this same approach.

Want to work with us? Have questions? Learn more below.

About The Role

As a Member of Technical Staff, Machine Learning Scientist on our Model Training team, you will be at the heart of our efforts to elevate model performance through innovative post training strategies. Your work will focus on optimizing our models after initial training by developing and implementing advanced fine-tuning methodologies. This role is critical to ensuring that our AI systems not only meet but exceed the demands of enterprise-scale applications.

This role is a strong fit if you:

• Bring deep experience training large-scale language models and building high-performance post-training pipelines.
• Are highly proficient in PyTorch and have expert-level knowledge of transformer-based architectures.
• Have a proven record applying fine-tuning methods like RLHF, DPO, and other reinforcement learning techniques to improve model behavior.
• Enjoy tackling complex engineering problems that directly influence model reliability, alignment, and performance at scale.
• Value practical impact and want to work closely with engineering and research teams to deploy cutting-edge systems in enterprise environments.

In this role, you will:

• Build and maintain scalable post-training infrastructure that incorporates advanced fine-tuning techniques.
• Research and implement new reinforcement learning approaches to improve model robustness, safety, and alignment.
• Collaborate across research, infrastructure, and product teams to productionize breakthroughs in model training.
• Run rigorous evaluations of model performance, iterating quickly based on experimental results and key metrics.
• Provide technical leadership on model optimization and training strategy, helping shape foundational AI systems for enterprise use.

language model training, RLHF, DPO, reinforcement learning, PyTorch, transformer architectures, post-training pipelines, model alignment, enterprise AI, model optimization, LLM fine-tuning, AI infrastructure, model robustness, scalable systems, AI safety