In a webinar titled "From production data to faster inference: the model optimization loop," Nebius explores the practical optimization loop in Token Factory, from capturing production logs and S3-backed datasets in Data Lab to transforming them into training and evaluation datasets. The session covers choosing optimization strategies such as supervised fine-tuning (SFT), distillation, or custom speculator training. Using speculative decoding as a concrete example, the webinar demonstrates how teams can turn real production data into workload-specific draft models and make informed trade-offs between latency, throughput, quality, and cost instead of optimizing for token price alone. Attendees will learn how production data becomes reusable training and evaluation data, where Data Lab, SFT, distillation, and custom speculator training fit into the optimization workflow, why speculative decoding is workload-dependent and how to determine if it is the right approach, and which metrics to measure before and after optimization to evaluate real production impact. The session is presented by Dylan Bristot, Head of Product Marketing, Token Factory; Mashrur Haider, Technical Product Manager; and Sujee Maniyam, Developer Advocate.