Large Language Models (LLMs) have the taken the world by storm, ever since ChatGPT was released in 2022. This has led to an increase in demand for LLM-related services.
Large Language Models are the best technology to manipulate and produce language. Hence, they are in demand in all different sectors.
From enhancing customer service through intelligent chatbots to revolutionizing content creation, and even aiding in complex data analysis, LLMs have proven to be indispensable tools. Their capacity to understand, generate, and interact using natural language has positioned them as pivotal assets in industries ranging from education and healthcare to finance and entertainment, driving an unprecedented demand for LLM-related services.
The Tesseract Advisory & Consulting team helped a client in the space of fintech develop its own LLM technology. In this case study we explore the why and the how.
Background
This particular client faced a challenge related to natural language processing. They were using an inefficient engine to translate the performance of financial portfolios into natural language.
The existing engine was fragile and prone to errors.
Our evaluation determined that the optimal solution would be to integrate a Large Language Model (LLM) with a Retrieval Augmented Generation (RAG) system. RAG is a specific kind of technology that bolsters the capabilities of LLMs by supplementing them with additional knowledge, thereby enabling the generation of higher quality answers to queries.
Solution – LLM, data augmentation & RAG
Our approach entailed refining a custom Large Language Model (LLM), complemented by a specialized Retrieval Augmented Generation (RAG) system to encode financial portfolio knowledge into a format accessible to the LLM.
To train the LLM, we employed data augmentation, utilizing another LLM to generate thousands of examples from a modest set of initial cases provided by a human expert.
The LLM delivered flawless results, as evaluated by a human expert.
The client now holds valuable, unique intellectual property, which they are leveraging to secure further funding.
Generative AI and LLMs
The success story with our fintech client underscores the transformative potential of Large Language Models (LLMs) and Retrieval Augmented Generation (RAG) systems within the financial sector. By customizing these cutting-edge technologies to address specific challenges, Tesseract has not only facilitated a leap in natural language processing capabilities but also empowered our client with a distinct competitive edge.
If you want to know more what we can do for your business, then make sure to get in touch. We specialise in delivering solutions in AI and data science, but also educating decision makers on topics such as data strategy and data maturity.
