Research Engineer, Post-Training (All Industry Levels)

As a Post-Training Research Engineer, you will advance alignment techniques and loss functions to improve data efficiency. You will design data pipelines to transform diverse web data into model-ready formats. You will identify quality signals to evaluate real-world performance and guide improvements. You will design sampling strategies to boost serving efficiency of large generative models. You will work across teams to implement GPU-accelerated training, distributed infrastructure, and robust, production-grade code. You will contribute to evaluating and deploying ML methods with a focus on reliability and scalability.