Graphlet AI

Graphlet AI is a consultancy working at the intersection of knowledge graphs and Large Language Models (LLMs). The two technologies are synergistic: LLMs that solve business problems are driven by knowledge bases best represented as graphs. We perform the following services: * Knowledge Graph Construction - we build large networks using natural language processing, information extraction, entity resolution and big data tools like Spark to model entire business domains and solve complex problems by applying graph machine learning [Graph ML] and artificial intelligence [Graph AI] to automate business processes and create new efficiencies. The networks we construct are in a form that is easy to use in a modern property graph or RDF database to drive prompt engineering or training custom LLM implementations. * Repeatable Entity resolution via Large Language Models [LLMs] - this technique allows us to encode nodes and edges as semi-structured data (JSON, XML) in text using a language model and then combine them based on semantic inferences made by the LLM, which has seen many similar documents as the nodes’ text representation on the world wide web. We made a 3 billion node business graph a 500 million node identity graph. * Training for Graphs and LLMs - it isn't enough to just hire a consultant... you need continuous bang for the buck. By training your staff to perform the work we do, we keep offering you value on an ongoing basis as your team takes over. Some things we've worked on include: LLMs, chatbots, graph machine learning, GNNs, PyGOD, visualizing large networks, NVIDIA RAPIDS, scalability, distributed systems, GPU computing, prototyping, data-driven web applications, proof of concepts, sales engineering, data science, data analysis, Elasticsearch, OpenSearch, vector search, information retrieval, Langchain, transformers, cybersecurity, intrusion detection, SIEMs, UEBAs, alert volume reduction I work in open source at https://github.com/Graphlet-AI/graphlet