name: Ketan Ayatti
role: DevOps Engineer Intern
company: Ionidea
study: BCA Final Semester
focus: CI/CD | Full Stack Apps | Infrastructure Automation | Linux Servers
mindset: "Build it right. Ship it reliably. Recover automatically."I'm Ketan Ayatti, a BCA Final Semester Student and DevOps Engineer Intern at Ionidea, working at the intersection of full-stack product engineering and modern infrastructure systems.
I don't just build applications β I take them all the way from first commit to a production-grade, containerized, automated deployment.
What I work on:
- Full-stack web platforms (React, Node.js, MongoDB, Socket.io)
- CI/CD pipeline automation (Jenkins, GitHub Actions)
- Docker containerization and multi-environment deployment
- Linux server infrastructure and NGINX configuration
- Resilient deployment strategies (Blue-Green, automated rollback)
- AI-integrated applications using Gemini, OpenAI, and Cohere
What I aim to ship:
β
Production-ready full-stack products
β
Automated and containerized deployments
β
Self-healing, scalable infrastructure
β
Systems that recover β not just systems that run
Development Foundations β Backend & Web Development (Node.js, Express, MongoDB)
Full Stack Products β React, Socket.io, JWT Auth, REST APIs
Containerization β Docker, Docker Compose, Multi-stage Builds
Infrastructure β Linux Server Deployments, NGINX Reverse Proxy
Automation β Jenkins CI/CD, GitHub Actions, Multi-branch Pipelines
Present β DevOps Internship @ Ionidea
Next β Kubernetes & Platform Engineering
Below are the systems I have built β each combining full-stack product engineering with DevOps-grade deployment practices.
A production-grade, full-stack communication platform that unifies real-time messaging, collaborative coding, secure file sharing, and AI-assisted workflows β deployed through an automated Jenkins CI/CD pipeline.
Repository: https://github.com/ketanayatti/communiatec-devops-platform
Full Stack Implementation:
| Layer | Technology |
|---|---|
| Frontend | React, Vite, Tailwind CSS, Zustand, React Router |
| Backend | Node.js, Express, Socket.io, Mongoose, JWT Auth |
| Database | MongoDB |
| AI & Media | Gemini API, Cloudinary |
| Auth | JWT + Cookie, GitHub OAuth, LinkedIn OAuth, RBAC |
Core Features Built:
- Real-time direct and group messaging with typing indicators, read receipts, and message reactions
- Collaborative coding sessions with multi-user cursor sync via dedicated Socket.io namespace
- Secure vault with owner-to-recipient file workflows and access enforcement
- Admin dashboard with user management, moderation, role updates, and audit logs
- Security-first architecture: hardened headers, CSP, input sanitization, HPP protection, rate limiting
DevOps Layer:
- Jenkins CI/CD pipeline with GitHub source integration
- Docker containerized deployment to Docker Hub registry
- Linux production server deployment with NGINX
An AI-powered mock interview platform with resume-based personalization, adaptive questioning, and multi-model evaluation β deployed through a branch-aware Jenkins CI/CD pipeline with Docker Compose.
Repository: https://github.com/ketanayatti/ai-mock-interview
Full Stack Implementation:
| Layer | Technology |
|---|---|
| Server | Node.js 20, Express 4, EJS views |
| Database | MongoDB 7, Mongoose |
| AI Models | Gemini (primary), OpenAI (optional), Cohere (optional) |
| File Handling | pdf-parse, mammoth, multer |
| Security | cookie-session, express-rate-limit, DOMPurify, CORS |
Core Features Built:
- Resume upload and extraction (PDF via
pdf-parse, DOCX viamammoth) - Gemini-powered resume-to-role summary generation
- Adaptive interview engine (10β15 questions based on experience level)
- Multi-AI evaluation pipeline across Gemini, OpenAI, and Cohere
- Performance dashboard with score trends and round-wise analytics
DevOps Layer:
- Multi-stage Docker build with non-root runtime user
- Branch-aware Jenkins pipeline:
developβ staging,mainβ production - Docker Compose files for dev and prod environments separately
- Automated build notification email in pipeline post-actions
A production-style deployment platform built around zero-downtime Blue-Green deployment with automated health validation and self-healing rollback.
Repository: https://github.com/ketanayatti/Autonomous-Self-Healing-Deployment-Platform
Engineering Focus:
- Blue-Green zero-downtime deployment strategy
- Automated health-check validation system
- Self-healing rollback mechanism on failure detection
- NGINX reverse proxy traffic switching between environments
- Docker-based container deployments
- Linux automation scripts for full infrastructure control
Engineering Goal: Design a resilient deployment platform capable of detecting application failures and automatically restoring service stability β without human intervention.
Central repository documenting DevOps experiments, architecture explanations, infrastructure designs, and deployment workflows.
Repository: https://github.com/ketanayatti/devops-engineering-portfolio
Developer Pushes Code
β
βΌ
GitHub
β
βΌ
CI/CD Pipeline (Jenkins / GitHub Actions)
β
ββββΊ Build & Test Stage
β
βΌ
Docker Image Build
β
βΌ
Container Registry (Docker Hub)
β
βΌ
Production Server (Linux + NGINX)
β
βΌ
Running Containers (Blue-Green / Rolling)
β
βΌ
Health Check + Auto Rollback if Failure
DevOps Engineer Intern β Ionidea
- Completed an end-to-end application deployment on a Linux-based server
- Worked with CI/CD automation workflows
- Practised infrastructure configuration, containerization, and deployment processes
product_thinking: ship features users care about
automation: mandatory β eliminate manual drift
security: built in from day one, not bolted on later
downtime: minimized through Blue-Green and health-check rollback
infrastructure: reproducible, containerized, environment-consistent
monitoring: important β you can't fix what you can't observe| Platform | Link |
|---|---|
| GitHub | https://github.com/ketanayatti |
| https://www.linkedin.com/in/ketanayatti/ | |
| Portfolio | https://ketanayatti.netlify.app/ |
"Great engineers don't build systems that never fail β
they build systems that recover automatically."


