MindMeld APP + AI Companion
MindMeld APP + AI Companion
Embedded Files
A cross-platform React application(16x9) deployed on Netlify (frontend) and AWS Elastic Beanstalk (Node.js backend) to deliver an AI-powered stress relief experience. Integrates OpenAI’s API for real-time, context-aware conversation with a virtual assistant that guides users through breathing exercises and supportive dialogue. Each session greets users with a dynamically fetched motivational quote. Features environment-secured API keys, CI/CD from GitHub, and responsive UI optimized for mobile and desktop. Designed for quick mental resets during high-stress moments.
A cross-platform React application(16x9) deployed on Netlify (frontend) and AWS Elastic Beanstalk (Node.js backend) to deliver an AI-powered stress relief experience. Integrates OpenAI’s API for real-time, context-aware conversation with a virtual assistant that guides users through breathing exercises and supportive dialogue. Each session greets users with a dynamically fetched motivational quote. Features environment-secured API keys, CI/CD from GitHub, and responsive UI optimized for mobile and desktop. Designed for quick mental resets during high-stress moments.
InboxAI, unfinished
InboxAI, unfinished
I’m building this because the AI surge makes it finally possible to solve one of the biggest daily problems: attention overload. Email is where people’s time, money, obligations, and decisions live, but today it’s still treated as a static inbox. This system connects to Gmail through a FastAPI backend, pulls full email data, and uses AI to understand it and then summarizing, prioritizing, extracting actions, identifying subscriptions, and drafting replies. The frontend becomes an intelligent workspace on top of Gmail, not just a reader. Instead of drowning in messages, users get structure, clarity, and saved time. This is exactly the kind of problem modern AI is meant to solve.
Image Processing
Image Processing
In this project, I developed an image processing engine in Java using a QuadTree data structure to represent grayscale images efficiently. Each pixel region was recursively subdivided only as needed, minimizing storage for uniform areas and enabling fast updates. I implemented all core components from scratch — including a dynamic circular queue for traversal and a recursive method to restructure the tree when pixel values changed. This structure allowed operations like brightness computation and image export to be performed directly on the tree, showcasing efficient spatial partitioning and memory-aware design.
In this project, I developed an image processing engine in Java using a QuadTree data structure to represent grayscale images efficiently. Each pixel region was recursively subdivided only as needed, minimizing storage for uniform areas and enabling fast updates. I implemented all core components from scratch — including a dynamic circular queue for traversal and a recursive method to restructure the tree when pixel values changed. This structure allowed operations like brightness computation and image export to be performed directly on the tree, showcasing efficient spatial partitioning and memory-aware design.
Suppose We Ran,
A 6x4 grayscale image was provided in PGM format:
P2
6 4
255
10 20 30 40 50 60
70 80 90 100 110 120
130 140 150 160 170 180
190 200 210 220 230 240
ResumeBoost AI
ResumeBoost AI
Live Site: https://resume-boost-ai-ogef.onrender.com. Developed a full-stack web application that intelligently analyzes resumes against job descriptions to generate personalized improvement suggestions. Users can upload .txt or .pdf files and receive tailored feedback powered by OpenAI's GPT-3.5 model. Built with Node.js, Express, and deployed on Render with automated CI/CD via GitHub. Integrated secure .env handling, file parsing (via pdf-parse), and dynamic prompt generation for actionable results.
Live Site: https://resume-boost-ai-ogef.onrender.com. Developed a full-stack web application that intelligently analyzes resumes against job descriptions to generate personalized improvement suggestions. Users can upload .txt or .pdf files and receive tailored feedback powered by OpenAI's GPT-3.5 model. Built with Node.js, Express, and deployed on Render with automated CI/CD via GitHub. Integrated secure .env handling, file parsing (via pdf-parse), and dynamic prompt generation for actionable results.
Reads the input provided by the user and suggests improvements based on the resume.
HIY Task Manager
HIY Task Manager
Built a fully functional Unix-style task manager in C, mimicking shell behavior with support for custom commands, process control, and I/O redirection.Implemented foreground/background execution, task suspension/resume with signals, and piping between tasks. Managed dynamic task tracking with state transitions and PID mapping. Used fork(), execv(), waitpid(), and sigaction() to handle multi-process concurrency, logging, and safe signal delivery. Robust, responsive, and signal-aware.
Built a fully functional Unix-style task manager in C, mimicking shell behavior with support for custom commands, process control, and I/O redirection.Implemented foreground/background execution, task suspension/resume with signals, and piping between tasks. Managed dynamic task tracking with state transitions and PID mapping. Used fork(), execv(), waitpid(), and sigaction() to handle multi-process concurrency, logging, and safe signal delivery. Robust, responsive, and signal-aware.
Clean command-line interface fully logged for real-time feedback.
A dynamic flow of Unix process states — Ready, Running, Suspended, Finished
Bellman-Ford
Bellman-Ford
Built an interactive GUI simulator to visualize and dynamically explore the Bellman-Ford shortest-path algorithm on complex directed graphs. Leveraged Java Collections and JUNG libraries to create a responsive, intuitive user interface for real-time graph editing and algorithmic step-through, enhancing user understanding of advanced graph concepts. - Vertices can be added, deleted, dragged, and connected freely.
Built an interactive GUI simulator to visualize and dynamically explore the Bellman-Ford shortest-path algorithm on complex directed graphs. Leveraged Java Collections and JUNG libraries to create a responsive, intuitive user interface for real-time graph editing and algorithmic step-through, enhancing user understanding of advanced graph concepts. - Vertices can be added, deleted, dragged, and connected freely.
Each Image shows Graph loaded, paths computed, another tested—negative cycle detected, simulation ends.
Tetris
Tetris
This project reimagines Tetris using custom dynamic arrays instead of Java collections, challenging the developer to build and manipulate 2D data structures from scratch. The gameplay includes rotation, flipping, and scaling of procedurally generated blocks, while the board updates dynamically based on player actions. The hardest parts include implementing dynamic resizing, handling irregular block transformations, and maintaining board integrity across edge cases—all without using Java's built-in data structures.
This project reimagines Tetris using custom dynamic arrays instead of Java collections, challenging the developer to build and manipulate 2D data structures from scratch. The gameplay includes rotation, flipping, and scaling of procedurally generated blocks, while the board updates dynamically based on player actions. The hardest parts include implementing dynamic resizing, handling irregular block transformations, and maintaining board integrity across edge cases—all without using Java's built-in data structures.
Suppose We Ran,
java Game 20 15
this would launch the game with a 20x15 dynamic board. -- All blocks are randomized.
Green Card Replacement System
Green Card Replacement System
Developed a Java-based GUI application to handle Green Card Replacement (Application) requests, integrating frontend data entry forms with backend validation logic. Leveraged UML modeling (class, sequence, activity diagrams to design system architecture, with componentized classes for name, address, and DOB validation. Followed Agile (Scrum) practices using GitHub and Maven for builds. Emphasized modular design, input validation, and real-time status updates.
Developed a Java-based GUI application to handle Green Card Replacement (Application) requests, integrating frontend data entry forms with backend validation logic. Leveraged UML modeling (class, sequence, activity diagrams to design system architecture, with componentized classes for name, address, and DOB validation. Followed Agile (Scrum) practices using GitHub and Maven for builds. Emphasized modular design, input validation, and real-time status updates.
Application Running Succecefuly after connecting to database with other part developed by teamates.
Page updated
Google Sites
Report abuse