CTTC empowers their customers with Generative AI services on AWS
ComeToTheClouds (CTTC) is a company specializing in AWS consulting services. Discover how privatesynapse.ai partnered with CTTC to seamlessly integrate Generative AI services on AWS.
Case Study
ComeToTheClouds approached privatesynapse.ai with a clear objective: to enhance their service offerings with cutting-edge Generative AI capabilities on AWS, while maintaining the highest standards of security and privacy. Cost-efficiency, performance, and scalability were identified as key priorities in their decision-making process for adopting AI solutions.
AWS has long been at the forefront of cloud technology, and in the era of AI, it continues to lead with innovative solutions. AWS introduced services such as Textract and Transcribe, which have seen widespread adoption. More recently, with the rapid growth of Generative AI and the emergence of large language models like ChatGPT, Llama, and others, AWS launched Bedrock—a platform designed to provide easy access to a variety of foundational models, along with integrated support for vector databases, prompt management, and agentic use cases. Among Bedrock's most notable offerings are the Sonnet models by Anthropic, developed through a strategic partnership, as well as leading open-source models such as Llama and Mistral.
During a strategic assessment to identify the best ROI opportunities for CTTC, we recognized AWS Bedrock as the optimal solution for enabling Generative AI capabilities natively within their AWS environment. We initiated a pilot project to set up the AWS environment for leveraging both LLMs and embedding models, focusing on a Retrieval Augmented Generation (RAG) use case.
First, we configured AWS Bedrock and enabled access to the following AI models: Embed 3 Multilingual by Cohere and Claude 3.5 Sonnet by Anthropic. To maximize throughput, we utilized Bedrock's cross-region inference capabilities.
Next, we deployed a vector database using Pinecone, integrated via the AWS Marketplace. Building on this foundation, we developed an API service using FastAPI and LangChain to create a Retrieval Augmented Generation (RAG) solution. This service enabled the ingestion of text documents into the Pinecone vector database, document embedding, and retrieval with LLM-powered question answering. The API was deployed using ECS Fargate, orchestrated with Terraform for scalable and reliable operations.
This project empowered CTTC with secure, private access to state-of-the-art Generative AI models, demonstrated through a practical RAG use case.
We strive to provide our customers with the latest of technology. Thanks to privatesynapse.ai, we are now able to leverage Generative AI on AWS.
Zoran Pajeska
CEO CTTC
