TNG Technology Consulting GmbH

We solve hard IT problems.

Latest Research

Check out our latest research:

• DeepSeek-TNG-R1T2-Chimera
  • Announcement on X
  • Model Card
• DeepSeek-R1T-Chimera
  • arXiv: Assembly of Experts
  • Announcement on X
  • Model Card
• Mixture of Tunable Experts
  • arXiv: Mixture of Tunable Experts
  • blog: Mixture of Tunable Experts

Blog

Read our latest blog posts:

• Prefill and Decode for Concurrent Requests - Optimizing LLM Performance
• Finetuning olmOCR to be a faithful OCR-Engine
• Efficient Request Queueing – Optimizing LLM Performance

Follow us

TNG on github , on X , on LinkedIn