The Problem
Our client, the creator of an AI-powered cybersecurity system, approached us to help bring their product to the public with a SaaS-like subscription model and Slack integration. Part of the system, however, was created with open-source code that needed cleaning up and securing. On top of this, our client also needed a static website to advertise their product.
Our Solution
After reviewing the client’s system and open-source libraries, we devised a plan to tighten up their application through advanced integrations with Amazon Web Services (AWS). We also applied Retrieval Augmented Generation (RAG) modeling techniques to enhance their AI technology further, and leveraged our expert knowledge of advanced software integrations to connect our client’s technology with Slack.
What We Did
RAG Backend
We deployed a backend on AWS to replicate the client’s RAG system using Hugging Face’s LLM. Our solution included training the model on client-provided PDFs, storing data in Cosmos DB, and integrating features like Chroma DB for VectorDB. In addition to providing API and Slack integration, we also logged user interactions and feedback.
Slack Integration
We created a Slack app that seamlessly connects to our RAG backend, allowing users to engage with Blodgic through their Slack workspaces. We then implemented a subscription/approval system for Slack access management in which users have limited free chats, with enhanced access granted post-subscription or upon admin approval. We also assisted in app approval and provided comprehensive documentation, including chat storage for administrative review.
Marketing Website
We designed a fully-responsive WordPress site with essential pages and critical product information. For this, a signup form with validation was integrated and linked to Mailchimp for seamless marketing integration. The site also features stock photographs chosen by the client, with content provided and managed collaboratively.