Project Details

This project aimed to build a self-healing infrastructure system capable of detecting failures in real time and automatically recovering services without manual intervention. The system was designed around the concept of proactive monitoring and automated response—ensuring maximum uptime and reliability in a containerized environment.

At the core of the project, Prometheus was used as the monitoring engine to continuously scrape metrics from a target service (in this case, an NGINX web server running inside a Docker container). The Blackbox Exporter was integrated to simulate HTTP probes and validate service availability. If the web server became unreachable, Prometheus would detect the probe failure based on the probe_success metric and trigger an alert.

Alertmanager was configured to receive these alerts from Prometheus and route them to a custom webhook endpoint. This webhook was built using a lightweight Flask API, acting as a bridge to execute automated tasks. Upon receiving the alert, the webhook would invoke an Ansible playbook designed to restart the NGINX container, thus healing the failed service and restoring normal operation—completing the feedback loop from monitoring to recovery.

This project not only strengthened my practical skills with monitoring and alerting tools but also demonstrated how DevOps automation can be leveraged to build reliable, production-ready systems. It showcased the real-world application of self-healing infrastructure, integrating multiple tools and scripts to create a resilient environment that minimizes downtime and manual overhead.

Project Details

This project was developed as part of a college semester project, with the primary objective of exploring modern web development technologies. In addition to fulfilling academic requirements, the project served as a hands-on opportunity to implement real-world development practices and enhance technical proficiency.

The core of the project was built using Node.js, which played a significant role in managing the user interface and overall functionality. Express.js was used to streamline server-side logic, while MongoDB served as the database for efficient data storage and management. Passport.js was integrated to handle user authentication, ensuring secure access and smooth authentication flow.

While HTML5 and CSS3 were used minimally for the landing page, the primary focus was on utilizing Bootstrap 4 and jQuery to enhance responsiveness and interactivity. These technologies helped create a seamless user experience while maintaining a clean and modern design.

By integrating these technologies, the project successfully delivers a full-stack web application that not only met academic objectives but also provided valuable experience in building dynamic and scalable applications. The hands-on approach reinforced key development skills and deepened the understanding of modern web technologies.

Project Details

NebulaBot is an AI-powered chatbot application for Android that leverages Google's Gemini API to deliver intelligent and real-time conversational experiences. Developed by adnanaiman360, the app is designed with a user-friendly interface for seamless interactions and dynamically supports Markdown formatting, allowing messages to be displayed with rich text elements like bold, italics, and code blocks.

It is likely integrated with Firebase for authentication and data management, ensuring a smooth user experience. Additionally, the app utilizes Android Jetpack components for optimized app architecture and efficient UI management. Built using Kotlin/Java, NebulaBot serves as a versatile AI assistant suitable for personal assistance, customer support, and educational purposes.

Its integration with the Gemini API ensures dynamic responses, making it a powerful chatbot solution for Android users looking for an interactive and well-formatted AI chat experience.

Project Details

The implementation of the AI-Based Customer Service Tool for Automated Warehouses marks a significant step toward integrating intelligent, document-aware automation into real-time support systems. By leveraging a Retrieval-Augmented Generation (RAG) model, the system is able to deliver both generic and context-specific responses with a high degree of fluency and factual accuracy. This is achieved through the seamless combination of document chunk retrieval using PineconeDB, PDF parsing via Langchain, and response generation through a lightweight FastAPI backend.

While the system currently addresses customer service within warehouse environments, its modular architecture makes it adaptable for various other document-driven domains such as healthcare, logistics, technical support, and education. Although some limitations were noted, particularly in handling ambiguous queries and response latency with very large datasets, these challenges offer clear directions for future development. Overall, the project successfully delivers a scalable, practical, and intelligent solution to modern customer service challenges.