Rapyder Empowers Stellaps with a Cutting-Edge Gen-AI Powered Multilingual Chatbot!

About the Client: Stellaps

Stellapps stands out as the first-of-its-kind startup in India dedicated to digitizing the dairy supply chain. Founded in 2011, it is an IIT Madras-incubated, Bangalore-based Internet of Things (IoT) startup with a primary focus on data acquisition and machine learning. Milk, the largest crop on this planet, highlights a strong demand for technology interventions, particularly in emerging markets where the yield per animal is low, traceability is inadequate, and quality is not up to the mark.

Industry:

FinTech

Offering:

GenAI Solution

AWS Services:

AWS VPC, NAT Gateway, Lambda, ECS with Fargate, ECR, AWS Bedrock, AWS RDS, API Gateway, S3

Business Need:

Stellapps needs a private chatbot to use its data for voice-based queries in multiple Indian languages. The chatbot should utilize GenAI and RAG techniques for relevant responses from unstructured documents. Several end users, farmers, and rural people use regional languages like Gujarati, Hindi, Kannada, Marathi, and Malayalam.

Below are the requirements from the customer:

1. Create a data ingestion solution to convert PDFs and CSVs into Vector format for storage in a Vector database, ensuring security and scalability.
2. Develop a secure and scalable API for speech recognition and administering English proficiency tests through audio inputs in various rural languages.
3. Develop a secure API to fetch answers from Vector databases based on user queries, utilizing large language models (LLMs) and Retrieval-Augmented Generation (RAG) techniques.
4. Develop a professional API to handle user audio inputs in rural languages, accurately translate them into English, and generate formatted answers from Vector database documents using LLM and RAG techniques, ensuring seamless response delivery.

Implementation:  

Our team of AWS-certified solution architects successfully implemented a microservice-based architecture on the AWS cloud for Stellapps. This deployment guarantees scalability, flexibility, and resilience, ensuring that the chatbot effectively meets Stellapps’s demands.

Our AWS-certified solution architects successfully implemented a microservice-based architecture on the AWS cloud for Stellapps. This deployment guarantees scalability, flexibility, and resilience, ensuring that the chatbot effectively meets Stellapps’s demands.

1. The AWS Mumbai region (ap-south-1) was introduced as a working region.
2. A separate network with a combination of VPC / Private Subnets was created according to best practices.
3. The proposal includes utilizing RDS DB with PostgreSQL PGvector for storing unstructured data as vector embeddings.
4. AWS S3 will be the storage solution for unstructured documents, managing the data ingestion into RDS via Lambda events.
5. AWS Bedrock models for Large Language Models (LLM) to generate answers using Langchain.AWS Lambda functions with an LLM agent to generate the answers and ingest the data.
6. Use API Gateway with Lambda functions to create a client-facing API for retrieving answers based on user queries.
7. AWS Lambda functions with an LLM agent to generate the answers and ingest the data.
8. Using ECS with Fargate to host a custom Flask API for the Whisper model, handling Automatic Speech Recognition (ASR) API tasks to convert audio to English transcription text.
9. ECR serves as the storage for the Docker file of the Whisper Custom Flask API.

Reaping Rewards:  

• Experienced a 25% surge in user traffic from rural areas due to enhanced multilingual accessibility.
• Serverless resources led to reduction in costs, maximizing efficiency.
• Reduced overall man-effort to process audios and multiple different formats within ecosystem