Case Study

DTEDI Smart Assistant - Autonomous AI Information Service

2025/Universitas Gadjah Mada (UGM)

DTEDI Smart Assistant is an AI-powered academic information service that operates autonomously around the clock. It is designed to instantly answer critical questions about the curriculum, administration, and facilities of DTEDI at UGM.

AWSDockerAIAutomationRAGLLMn8n
DTEDI Smart Assistant - Autonomous AI Information Service preview

What I Worked On

My Responsibilities

CL

System Design & Cloud Infrastructure

Designed the end-to-end system architecture and spearheaded AWS EC2 deployment with Docker and n8n.

RE

Automated Knowledge Ingestion

Built an n8n workflow that detects Google Drive updates, downloads documents, splits text, and updates Pinecone.

ZA

Live RAG Pipeline

Designed a real-time retrieval-augmented generation pipeline using Pinecone, Gemini, and webhooks.

GL

Multilingual Semantic Understanding

Integrated Cohere multilingual embeddings for formal and informal Bahasa Indonesia queries.

ME

Conversational Memory

Implemented LangChain memory buffer through n8n for coherent multi-turn conversations.

Stack & Links

Built With

Tools, frameworks, and platforms used to ship this project.

AEAWS EC2
Docker
n8n
Google Gemini
LangChain
PPinecone
CCohere
Google Drive API

Project Links

Screenshots

Project preview carousel

DTEDI Smart Assistant - Autonomous AI Information Service screenshot 1

Screenshot 1 / 4