A scalable, enterprise-ready RAG (Retrieval-Augmented Generation) platform designed to accelerate onboarding and engineering productivity. This multimodal assistant allows developers to ask questions about their codebase, documentation, Slack threads, Confluence pages, and screenshots—all from a single unified interface.
- Multimodal RAG Architecture: Combines text and image understanding via LangChain, FAISS, and Gemini Pro.
- Live Slack & Confluence Integration: Continuously ingests and indexes Slack messages, Confluence pages, and repository content.
- Vision-Aware Querying: Supports screenshot-based answers using OCR and Gemini Vision models.
- Contextual Responses: Displays answers with reference to source (code/Slack/image/Confluence).
- Streamlit UI: Intuitive web interface to interact with the assistant.
- GPU Support: CUDA-compatible backend built for GKE GPU node pools.
| Layer | Tech |
|---|---|
| Backend | Python, FastAPI, LangChain, FAISS |
| Frontend | Streamlit |
| ML Models | Gemini 1.5 Pro (Vision + Text), SBERT |
| Storage | ChromaDB / FAISS (Vector Store) |
| Cloud | Google Cloud Platform (GKE, Container Registry) |
| DevOps | Docker, Kubernetes, GitHub Actions |
| Integrations | Slack API, Confluence API |
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.backend.txt
- Embed Documents
python ingest/embed_docs.py- Push Docker images
docker push juniemariam/kt-backend:latest
docker push juniemariam/kt-frontend:latest- Apply K8s YAMLs
kubectl apply -f k8s/backend-deployment.yaml
kubectl apply -f k8s/backend-service.yaml
kubectl apply -f k8s/ui-deployment.yaml
kubectl apply -f k8s/ui-service.yaml