At Flink Labs, our research fuels innovation. We undertake ongoing applied R & D into A.I., complex systems, and artificial life, transforming cutting-edge theories into practical applications. Our relentless curiosity and commitment to exploring new technology drive us to deliver groundbreaking solutions.
Artificial intelligence plays a role in almost all out projects, automating manual processes to deliver immediate productivity gains and as a foundation for the creation of new A.I. native products.
We have used machine learning algorithms to cluster and visualise data; deep neural networks such as fine tuned BERT models to embed and classify hundreds of thousands of pieces of customer feedback; language models and custom natural language pipelines for automated narrative generation to produce textual descriptions of data in reports; image recognition to assist in patent analysis; and markvov chains for predicting customer movement through supermarkets.
We are actively working on projects involving generative A.I. and A.I. alignment focusing on mechanistic interpretability.
We use LLMs within many live client engagements for text classification, generation of reporting content, and real time user question/answering. Our custom fine-tuned BERT model is used to classify feedback from hospital patients. We use GPT models to assist in the generation of content for reporting and dashboards. We use chat models to orchestrate responses to user queries and data exploration. And we augment LLMs with private client data and embeddings to provide exploratory question/answering interfaces.
We have active internal projects using GPT-4 for agentic problem solving and have built custom language models to assist in explaining decisions made by A.I. systems. We are also undertaking internal research on A.I. safety and alignment, specifically focusing on mechanistic interpretability.
We use algorithms inspired by natural biological phenomena (such as the flocking of birds, schooling of fish and ants foraging) to help our clients understand the complex dynamics they operate in, to predict future scenarios and to navigate the complex and uncertain path of our rapidly evolving future.
We use tools such as agent based modelling, digital simulations and complexity science to gain insights into digital ecosystems and processes, helping us to create more robust, adaptive products.
Our team attended the prestigious Complex Systems Summer School at the Santa Fe Institute in 2022. Flink Labs is a member of the International Society for Artificial Life.