GenAI Accelerators
Talk to us now!
Solve challenging business problems using our accelerators & start your analytics journey with a strong and confident foundation.
VartikGPT Secure Extensible Enterprise Chat
Chat based accelerator that have generic Data Ingestion pipeline to ingest unstructured and semi-structured data, chunk using multi chunking strategies like Recursive, Semantic and store in Vector database. It uses open-source embedding framework to embed text. It supports using multiple LLMs including open source, caching, chat history, routing for cost optimization. It can be extended to domains like HR, Finance, IT through configurable Prompts and separate Vector database indexes.
Generic Data Ingestion Framework
Enterprise ready secure and scalable data ingestion framework ingest data from private knowledge bases and support multiple chunking strategies and Vector databases like Pinecone, Qdrant, Azure AI Search. It allows using multiple LLMs including open source LLMs for text embedding. It ingest PDF, Docx, PPT, mails, CSV, MD data formats and many more.
Dynamic Document Generation
Configurable and metadata driven document chunk generation Accelerator that uses multi level Dynamic Generation Map to streamline document chunk dependencies. It uses configurable Prompt definitions in YAML files in datastore. It uses Aress Generic Data Ingestion Framework to ingest historical data for contextual reference for document generation. This accelerator can be used as reusable framework for new use cases.