Back to projects
◈ Architecture AWS

TPQI AI Qualification Intelligence Platform

AWS-native RAG + Graph AI for Thailand's Professional Qualification Institute

Full-stack AI platform for สถาบันคุณวุฒิวิชาชีพ (TPQI). Ingests qualification documents via OCR workers, builds a graph of job-role → skill → competency relationships in Neo4j/Neptune, and serves RAG-powered queries through ECS Fargate. Gemini AI (Vertex AI) handles inference with cost guardrails.

✓ Verified StackNotion: Architecture diagram embedded in NT-POC page, Jan 2026
TRY THE WORK

The system is deployed on edge nodes. Monitor container logs, memory allocations, and trigger virtual requests below:

edge-console://tpqi-platform.dev
LIVE PREVIEW

Resource Monitor

EDGE NODE CPU12%
RAM ALLOCATION18.4MB / 128MB
PLATFORM CLUSTERAWS
KEY METRICS & FACTS
AWS
Cloud platform
ECS Fargate (serverless containers)
Compute
SQS (OCR + Inference workers)
Async queue
pgvector / OpenSearch
Vector store
Neo4j / AWS Neptune
Graph database
Gemini via Google Vertex AI
LLM
Cognito + IdP SSO + RBAC
Auth
OTel → CloudWatch
Observability
CloudFront + WAF
CDN