🛰️ Daily AI Frontier
‹ back to 2026-09-02

From Production Traffic to Post-Training: Building a Self-Hosted LLM That Covers the Corporate Request Mix

Research LLMs & Foundation Models

Ranking

Overall 86
Content 95
Popularity 66

Observed public metrics from 1 member.

Representative image for From Production Traffic to Post-Training: Building a Self-Hosted LLM That Covers the Corporate Request Mix

Merged summary

TL;DR - This paper presents a post-training recipe for consolidating traffic from more than 200 enterprise applications onto one self-hosted LLM. Specialized GRPO experts and model merging close production-specific quality gaps while reducing GPU fragmentation and serving costs.

  • Production errors are grouped into instruction following, function calling, and internal task-distribution gaps, with evaluations stratified to real traffic.
  • Separate GRPO experts avoid cross-domain reward interference and are combined using two-stage SLERP.
  • Domain-specific fixes address semantic collapse, excessive function calls, and reward-driven verbosity.
  • The resulting model reportedly beats a roughly 7Ă— larger baseline on internal Arena, instruction-following, and function-calling evaluations, and serves 116 million monthly requests—50% of platform traffic.

Sources (1)

From Production Traffic to Post-Training: Building a Self-Hosted LLM That Covers the Corporate Request Mix

arXiv cs.CL Olga Tsymboi, Dmitrii Stoianov, Ramil Latypov, Danil Taranets, Daniil Dryabin, Mikhail Gashkov, Viktor Zelenkovskiy, Aleksandr Fida, Gleb Alektorov, Nikita Gulyakov, Arthur Babkin, Aleksandr Medvedev, Pavel Gein, Anatolii Potapov 2026-09-01 arXiv:2609.01572
Public signals Hugging Face upvotes 35
Providers: Hugging Face · Upvotes 35 OpenAlex · N/A Publisher · N/A Semantic Scholar · N/A X · N/A Fetched 2026-09-25 14:24:56.807137 UTC

TL;DR - This paper presents a post-training recipe for consolidating traffic from more than 200 enterprise applications onto one self-hosted LLM. Specialized GRPO experts and model merging close production-specific quality gaps while reducing GPU fragmentation and serving costs.

  • Production errors are grouped into instruction following, function calling, and internal task-distribution gaps, with evaluations stratified to real traffic.
  • Separate GRPO experts avoid cross-domain reward interference and are combined using two-stage SLERP.
  • Domain-specific fixes address semantic collapse, excessive function calls, and reward-driven verbosity.
  • The resulting model reportedly beats a roughly 7Ă— larger baseline on internal Arena, instruction-following, and function-calling evaluations, and serves 116 million monthly requests—50% of platform traffic.
item →