Kevin Tan, Cloud Solutions Architect

About Me

I am a Singapore-based cloud architect and engineering leader with 17+ years of experience building enterprise systems that scale and continue operating when things go wrong. My career spans high-availability platforms, distributed systems, and cloud-native architectures in mission-critical environments. I believe the best leaders stay hands-on, close to the code, the architecture, and the people. Recently, I have been exploring how AI agents can safely interact with real enterprise systems, moving beyond demos into production-ready integrations.

Writing

I write about production AI agents, MCP servers, and cloud infrastructure at blog.jztan.com.

Claude Code PDF Limit: How to Read Large PDFs (pdf-mcp)

Why Claude Code can't read large PDFs, and the MCP server I built to fix it.

The Production AI Agent Playbook

Eight disciplines for building LLM systems that actually hold up under load.

AI Agents Fail Silently: 5 Error Handling Patterns for Production

What goes wrong in production, and the patterns I use to catch it.

How to Ship an MCP Server to Production

Authentication, safety, and everything that broke before it worked.

AWS Global to China S3 Sync: A Serverless Solution

The cross-region sync problem that broke us, and the serverless fix.

Read all posts at blog.jztan.com →

Experience

Deputy R&D Director

GDC Technology • Jan 2022 - Present

  • • Directed architecture and industry acceptance of Digital Cinema Enterprise Software Solutions, sustaining GDC leadership in the digital cinema space.

  • • Led and mentored a distributed team of more than 10 engineers across two regions, improving delivery efficiency by 20 percent and developing future senior leaders.

  • • Architected and owned AWS serverless service platforms supporting mission-critical operations, achieving 50 percent scalability improvement and reducing operational overhead by 40 percent.

R&D Management Roles

GDC Technology • Jan 2013 - Dec 2021

Titles held: Senior R&D Manager, R&D Manager, Deputy R&D Manager

  • • Built and scaled cross-functional engineering teams during periods of rapid organizational growth.

  • • Acted as technical liaison with industry standards bodies and strategic customers, aligning product direction with evolving requirements.

  • • Led design and delivery of second-generation Theatre Management System (Cinema Automation 2.0), modernizing cinema operations.

  • • Introduced modern SDLC practices including Git-Flow, JIRA, Confluence, and Kanban, reducing development cycles by 30 percent.

R&D Engineering Roles

GDC Technology • Jul 2007 - Dec 2012

Titles held: Senior R&D Engineer, R&D Engineer

  • • Designed and delivered first-generation Theatre Management System deployed across large cineplex environments.

  • • Developed electronic cinema playback systems with HDCP security compliance and high reliability requirements.

Software Engineer

Cyberlink • Jun 2006 - May 2007

  • • Developed multimedia software solutions in Python, C++, and DirectShow for global OEM partners including HP, Dell, and NEC.

Education

National Taiwan University

Bachelor of Science (BSc) Computer Science • Jun 2006

Certifications

AWS Certified Solutions Architect - Associate
Issued April 2024 · Expires April 2027

Technologies

Experienced engineering leader with deep expertise in cloud architecture, enterprise SaaS platforms, and modern software development practices. Strong hands-on background in AWS, serverless computing, distributed systems, and AI integration. Proven track record leading teams, modernizing legacy platforms, and delivering reliable systems in mission-critical environments. I focus on building technology that is practical, scalable, and operationally sound.

Skills & Tools

Core Strengths:

AWS Architecture Serverless Solutions IaC (CloudFormation) Security & Compliance Cost Optimization Distributed Systems Observability AI Integration Developer Tooling API Design
Software Development
Python Django C/C++ PostgreSQL API Design & Integration
Cloud & DevOps
AWS Serverless Framework (IaC) CI/CD (GitHub Actions)
Practices
Agile Methodologies/Kanban AWS Well-Architected Framework Performance Tuning
Domains
Digital Cinema/DCI/SMPTE Software Product Management
AI & Integration
Model Context Protocol (MCP) FastMCP Framework AI Agent Integration

Patents

System and method for providing failover of storage for digital cinema broadcasting
US 10,798,427 B2 · Issued: October 6, 2020
System and method for cinema play memory failover
CN 107743245 B · Issued: July 24, 2020

Portfolio

blueclaw

AI agent observability

Summary:

Open-source CLI for AI agent observability, context masking, regression testing, and HTTP API. v2.0+, published on PyPI.

Highlights:

  • • CLI tool for tracing and observing AI agent behaviour in production.

  • • Context masking for safe logging of sensitive data.

  • • Regression testing support and HTTP API for integration.

Tech Stack:

Python CLI Observability PyPI

pdf-mcp

MCP server

Summary:

MCP server that lets Claude Code and AI agents read PDFs beyond their token limits. Strong community traction.

Highlights:

  • • Solves Claude Code's large-PDF token limit with a lightweight MCP server.

  • • Built on FastMCP; published on PyPI with strong community adoption.

  • • Used in production AI agent workflows.

Tech Stack:

Python MCP FastMCP PyPI

Redmine MCP Server

AI Integration & Developer Tools

Summary:

Production-ready Model Context Protocol server enabling AI assistants to interact with enterprise Redmine systems with strong security and operational guarantees.

Highlights:

  • • Built production-ready MCP server using FastMCP, enabling AI agents to manage enterprise project workflows programmatically.

  • • Implemented secure authentication supporting API keys and credential-based access with automatic session handling.

  • • Designed modular architecture with testing and Docker-based deployment for scalable production use.

  • • Implemented secure file handling with automatic expiry and cleanup for enterprise compliance.

  • • Published to PyPI with full documentation to support community adoption.

Tech Stack:

Python FastMCP MCP Docker Redmine API PyPI Security

Get in touch

I write and build in public about AI agents, MCP, and production cloud systems. I'm always happy to hear from fellow practitioners, and occasionally open to advisory conversations or speaking opportunities.

LinkedIn GitHub Medium Blog