Tian Tan

• AWS Certified Solutions Architect – Associate | DevOps & Full-Stack Developer. Highly motivated developer with a proven track record of architecting and deploying scalable, secure cloud environments. AWS Certified (SAA-C03) with hands-on expertise in Infrastructure as Code (Terraform), containerization (Docker), and automated CI/CD pipelines (GitHub Actions). Experienced in full-stack development with Next.js and TypeScript, focusing on high-availability system design and secure identity management. Adept at technical discovery and codebase analysis, with a deep interest in optimizing the developer experience through robust automation and zero-downtime deployment strategies.

AI-Powered Talent Discovery App: Next.js, AWS Cognito, DynamoDB, Terraform, Docker, GitHub Actions, Google Gemini API

• https://talentdiscovery.xyz/
• Architected Cloud Infrastructure: Provisioned a full AWS stack including VPC, EC2, DynamoDB, and Cognito using Terraform for reproducible, version-controlled Infrastructure as Code (IaC).
• Integrated AI Assessment Engine: Developed five core assessment features utilizing the Google Gemini API with structured JSON outputs and real-time streaming for a responsive user experience.
• Secured CI/CD Pipelines: Built an automated workflow using GitHub Actions and OIDC federation, leveraging IAM roles via AssumeRoleWithWebIdentity to eliminate stored AWS credentials.
• Optimized Container Deployment: Orchestrated a Docker-based deployment on EC2 behind an Nginx reverse proxy with Let’s Encrypt SSL, featuring a zero-downtime container swap mechanism.
• Advanced Secret Management: Enforced a "zero-secrets-in-code" policy by managing all application sensitive data and API keys through AWS SSM Parameter Store.
• Engineered Network Isolation: Configured custom VPC networking and IAM roles following the principle of least privilege to secure data flow between the application and database layers.

NutriTrack- AI-powered Nutrition SaaS: Next.js, TypeScript, AWS (EC2, ECR), Docker, Terraform, GitHub Actions, Stripe, Clerk

• http://54.252.116.221/
• Full-Stack SaaS Architecture: Developed a production-ready platform integrating Stripe for subscription billing, Clerk for authentication, and AI-driven meal analysis via Gemini 2.0 Flash.
• Provisioned AWS infrastructure as code using Terraform — EC2, ECR, Elastic IP, IAM roles with least-privilege ECR pull permissions
• Test-Driven Reliability: Authored 47 unit tests with Vitest to validate business logic, successfully identifying and resolving a critical edge-case bug in the user activity streak calculation.
• Streamlined Containerization: Utilized multi-stage Docker builds to minimize production image sizes and automated secret injection via GitHub Secrets for secure, manual-free deployments.
• Automated CI/CD Workflows: Engineered a GitHub Actions pipeline with parallel linting, type-checking, and unit testing jobs to gate all deployments to the AWS environment.
• Resolved Production Bottlenecks: Debugged and fixed complex infrastructure issues including 502 gateway errors, SSH host key conflicts, and Clerk authentication redirect loops.

End-to-End GitOps Microservices Pipeline AWS (EKS, EC2, VPC), Terraform, Kubernetes, Docker, Argo CD, GitHub Actions, Route53

• Automated Infrastructure as Code: Architected a production-ready AWS environment (VPC, EKS Cluster, IAM) using Terraform modules, implementing remote state management and state locking for environment consistency.
• Microservices Containerization: Engineered a multi-stage Docker workflow to containerize a Java-based microservice architecture, optimizing image security and size before pushing to Amazon ECR.
• Kubernetes Orchestration & Networking: Orchestrated high-availability deployments on Amazon EKS; configured Load Balancers, Ingress Controllers, and Route53 DNS mapping to securely expose the frontend to external users.
• Declarative GitOps CI/CD: Established a full GitOps lifecycle by integrating GitHub Actions for Continuous Integration and Argo CD for Continuous Deployment, achieving single-touch, self-healing automation from code to cluster.