JOBLOOP Intelligent Job Discovery Engine

JobLoop is a full-stack job discovery platform designed to collect real-time job listings directly from company career pages instead of relying on centralized job feeds. The system initially seeds itself from platforms like Peerlist and Y Combinator, then expands organically across the web. The backend is built using Go and Playwright, where I developed an adaptive automation engine that dynamically understands page structure, accessibility attributes (ARIA labels), pagination controls, and popular ATS platforms like Greenhouse, Lever, Workday, and Ashby.

Rather than hardcoding pagination patterns, JobLoop performs a discovery pass on each site to learn how job navigation works, then safely iterates pages until data is exhausted. The system supports URL-based pagination, click-based navigation, iframe ATS portals, and JavaScript-driven layouts. Using this approach, JobLoop can easily aggregate 5000+ jobs — exceeding the coverage of many traditional job aggregators.

A unique aspect of JobLoop is its recursive discovery model. After scraping a company, the system extracts testimonials, converts testimonial images back into text, identifies referenced companies, and automatically searches those companies for jobs as well. Each discovered company feeds back into the system, creating a continuous loop that expands the job graph without manual input.

The frontend is built with React, providing a clean interface for browsing opportunities aggregated directly from source companies. Unlike traditional job platforms, JobLoop follows a decentralized approach — any company website can become a data source automatically, resulting in fresher listings and fewer duplicates. This project highlights my experience with Go, Playwright, React, Browser Automation, Pagination Intelligence, OCR, Recursive Crawling, and Full-Stack System Design, showcasing production-grade automation built for real-world web complexity.

https://jobloop.dev

Backend Source Code

Frontend Source Code

BANK-OF-ANTHOS

To automate the development lifecycle, I implemented a CI/CD pipeline with Jenkins in the application build repository. This pipeline handled code compilation, unit testing, and Docker image creation, pushing updates to Docker Hub for seamless deployment. This automation streamlined processes, ensuring quality and timely delivery—key aspects of my engineering approach.

For deployment, I used Kubernetes on AWS, managed with ArgoCD for GitOps automation. The "bank-of-anthos-k8" repository contains manifests I configured using kustomize's base and overlays structure, enabling modular setups for different environments. ArgoCD synced the cluster with the repository, while ArgoCD-image-updater automatically updated Docker image tags from Docker Hub, keeping deployments current with minimal effort. I set up the AWS infrastructure using Terraform, provisioning resources like EC2 instances and networking, likely via Elastic Kubernetes Service (EKS), to ensure scalability and security.

This project reflects my ability to integrate Node.js, Java, Python, Jenkins, Terraform, ArgoCD, ArgoCD-image-updater, Building Docker images using Maven, Testing Docker Images and kustomize into a cohesive, AWS-hosted solution. It highlights my expertise in building scalable, automated systems that balance innovation and reliability, positioning me to tackle complex challenges in software engineering.

See Project Source Code

INSTA-TWITTER AUTOMATION

The Twitter-Insta-Bot is a browser extension that automates uploading tweet media across my Instagram Account. I dreamt of hav ing a simple script that excutes downloading the tweet media and posting it to my instagram account upon click of an icon which was appended on every tweet, this extension leverages a Node Js tech stack including Puppet for browser automation, twitter API for media download. It is built with an idea of any individual could increase his/her social media visibility, this tool offers a flexible solution.

Node.js, JS, Puppeter, ExpressJs, Twitter API, Tailwind CSS powers this automation extension

See Project Source Code

IMAGE STORE

Image Store is a dynamic platform that enables users to upload images, which can then be viewed, liked, and commented on by other connected users. To ensure secure access, the platform incorporates robust authentication mechanisms, restricting access to private routes and features only to authenticated users.

This project is built using modern web development technologies, ensuring a responsive and user-friendly interface. The authentication system is designed to protect user data and maintain the integrity of user interactions within the platform. By combining image sharing capabilities with social interaction features, the image-store platform offers a comprehensive solution for users seeking a secure and engaging environment to share and discuss visual content.

This web Application is powered with NextJs, Reactjs, MongoDB, Cloudinary, Redux Toolkit, Tailwind CSS. Handles Seamless Authentication system with username and password. The application is deployed on Vercel and the images are stored in Cloudinary. The application is designed to be responsive and user-friendly, ensuring a seamless experience for users across devices. The platform offers a secure and engaging environment for users to share and interact with visual content.

See Deployed Version

See Project Source Code

ChartJs Bar Chart

I have created a simple bar chart using ChartJs, to implement the sorting functionality visually on the chart. The chart is designed to display the data in a visually appealing way, making it easy for users to understand the information at a glance. The sorting functionality allows users to rearrange the data based on different criteria, providing a more interactive and engaging experience.

See Deployed version

See Source Code

My Three.js Personal Portfolio

The website you are currently reading, utilizes blender for the room model, three.js for rendering and camera movements, GSAP for high quality animations, and css/html.

I had originally tried to create an interactive room using Spline 3d, in which the camera would guide the user experience like virtual reality. I had to give up on using their platform as it would not let me edit their source code after designing the room. After being inspired by the blender room design I had found, I utilized that instead and used three.js to create the website you are currently on.

See Deployed version

See Source Code