Efficient training of machine learning algorithms

Optimisation of results at reduced costs

During this year's GenAI DACH conference in Berlin, it was a pleasure to deliver a presentation that spanned various topics, including Huawei Pangu-Weather, end-to-end lifecycle implementation in AI projects, fine-tuning with LoRA, and an introduction to Retrieval-augmented generation (RAG) system.

RAG systems, in particular, allow customisable application architectures using large-language models (LLMs). These systems are noteworthy for their accessibility, flexibility in determining architectural components' deployment, and ability to support global architectures with local adaptations. Furthermore, even partial implementations, such as employing a vector database for semantic search, can offer significant value.

Download: https://shrnk.eu/ai5vs