←SELECTED WORKS•CASE 04•automation

AI-Powered DAM Metadata Uploader

Automated AI ingestion pipeline and Model Context Protocol (MCP) server for CUHK Canto Digital Asset Management, featuring 5-skill agent, spatial college derivation, and binary IPTC exiftool embedding

CANTO DAM AI INGESTIONMCP SERVER ENABLED
canto-uploader.xavinno.com // mcp
Canto DAM Ingestion Pipeline & MCP Server
Modular 5-skill agent on Vertex AI with spatial college derivation, exiftool IPTC embedding, and remote MCP tools.● LIVE
✦ Pipeline: 5 Skills (agent.md) + Spatial College Derivation + S3 Polling + MCPCONNECTED
SYSTEM // CANTO_DAM

Enterprise DAM Ingestion & MCP Server

5-skill autonomous agent deriving spatial college taxonomy with binary IPTC injection and remote MCP tools.

Canto DAM Ingestion Pipeline & MCP Server

Modular 5-skill agent on Vertex AI with spatial college derivation, exiftool IPTC embedding, and remote MCP tools.

01 // PROVEN IMPACTVERIFIED
0+ Buildings

Automated landmark identification across 170+ buildings and 40+ landmarks with spatial college derivation.

MCP ENDPOINT: canto-uploader.xavinno.com/mcp● 10 ALBUMS AUTOMATED
02 // THE CONTEXTPROBLEM ➔ FIX
THE CHALLENGE

Manually cross-referencing 170+ buildings, 40+ landmarks, and 10 album categories made campus photo ingestion error-prone.

THE SOLUTION

Modular 5-skill AI agent on Vertex AI with spatial majority-vote college derivation and remote MCP tools for AI clients.

ZERO LONG PARAGRAPHS
WORKFLOW // CANTO DAM ↔ AI UPLOADER ↔ VERTEX AI PIPELINE
Canto DAMPhotos + metadataAI UploaderCentral hubEXIF + capture dateCollege lookup tableEnrich → sync to CantoGoogle Vertex AIRead photos + skillsBuilding · tagsDescriptionAPI →← DataPhotos →← TagsPhoto queueXavier’s 5 SkillsGuide image analysisUploadInstructions ↑
CANTO REST API ↔ CENTRAL HUB ↔ VERTEX AI · 5 SKILLSEXIF & COLLEGE MATRIX SYNCED ✓
02 // XAVIER'S ROLE & HOW HE BUILT ITLEAD INTEGRATION ENGINEER & AI AGENT ARCHITECT

CASE STUDY ROLE: LEAD INTEGRATION ENGINEER & AI AGENT ARCHITECT

Xavier designed the modular 5-skill AI agent where each tagging instruction (albums, buildings, landmarks, descriptions, tags) is stored in a clean markdown file.

He developed a spatial majority-vote algorithm that deduces which of the 9 colleges a photo belongs to, used exiftool to stamp IPTC metadata into the image binaries, and built an MCP server so AI assistants like Claude can search and upload photos using plain English.

WHAT XAVIER DIRECTLY DELIVERED & SHIPPED
Designed a modular 5-skill AI agent (agent.md + skills) that inspects 170+ buildings and 40+ landmarks in one pass
Created a spatial majority-vote algorithm that automatically derives which of CUHK's 9 colleges a photo belongs to
Used exiftool to permanently write IPTC keywords and descriptions into image files before upload
Built a resilient S3 upload and polling backend that updates Canto's Batch Edit APIs automatically
Developed an MCP server allowing AI assistants (Claude Desktop, Cursor) to search and upload to Canto via chat
HANDS-ON OWNERSHIPDIRECT ARCHITECT & BUILDER
03 // LANGUAGES & TOOLS USED
LANGUAGES & CORE RUNTIME
JavaScript (Node.js)TypeScriptMarkdown (Prompt Skills)HTML5
FRAMEWORKS & ENGINES
Express.jsexiftool-vendoredexifrDocker Compose
TOOLS, AI MODELS & APIS
Google Vertex AI (Gemini 2.5)Canto REST API v1AWS S3 UploadModel Context Protocol (MCP)
STORAGE & INFRASTRUCTURE
Docker StackCloudflare Zero Trust TunnelSynology NAS / VPS
CASE STUDY TECH ECOSYSTEM
Node.jsExpress.jsGoogle Vertex AICanto Batch Edit APIexiftool-vendoredexifrAWS S3 UploadModel Context Protocol (MCP)Docker
TECH ARCHITECTURE● FULLY FUNCTIONAL
04 // HOW THE SYSTEM WORKS

HOW THE CANTO UPLOADER WORKS

4-STEP PIPELINE
STEP 01
01/04

Drop Photos or MCP Prompt

User drops photos into the web portal or instructs Claude via natural language MCP tools.

StatusActive
STEP 02
02/04

5-Skill AI Analysis

Gemini identifies 170+ buildings, 40+ landmarks, 10 albums, and writes 20 search tags.

StatusActive
STEP 03
03/04

College & EXIF Stamping

Deduces college affiliation via majority vote and embeds permanent IPTC metadata via exiftool.

StatusActive
STEP 04
04/04

Canto S3 & Batch Edit

Uploads binary to Canto's S3 CDN, polls for asset creation, and fills all metadata fields.

StatusActive