Talk "Tales from the machinery room - Customizing LLMs @ TNG AI research"

Way beyond this X-post, the R1T and R1T2 Chimera models published by TNG gained severe attention with a daily usage of more than 10 billion tokens via OpenRouter. So, what kind of "Frankensteining" is going on there? How can a small software consultancy such as TNG produce its own models?

At TNG, we, an internal AI research group, have been experimenting and publishing results with a focus on Mixture-of-Expert Large Language Models and adaptions of the DeepSeek model family. First, we manipulated the way experts work within a model under the name "Mixture of Tunable Experts". We then continued with the "Assembly-of-Experts" merging process resulting in the Chimera models.

This talk by my TNG colleague Fabian Klemm and me at TNG’s Big Techday on October 24, 2025, aimed to give some technical insights into the work of TNG AI research. We recalled the theoretical basics and the most important results. Furthermore, we provided technical details on how things were done as well as some anecdotes about the successes and losses along this journey.

You can find a video recording of the talk on YouTube.