ML Engineer & Full Stack Developer | Ex-SWE Intern @Dell Technologies | Final year B.Tech CSE AI & ML @UPES, India

projects

github

leetcode

codeforces

Amulya Jain • Peerlist

<h3>Features</h3><ul><li>Single Player Mode: Play against an intelligent AI opponent using minimax algorithm</li><li>Multiplayer Mode: Two players can play against each other</li><li>Scoreboard: Track and view game history and outcomes</li><li>Data Analysis: Visualize win percentages with pie charts using matplotlib</li><li>GUI Interface: Clean and intuitive graphical user interface using Tkinter</li></ul><h3>How to Play</h3><ol><li>Launch the Game: Run TicTacToe.py to open the main menu</li><li>Choose Game Mode:<ul><li>Single Player: Play against the computer AI</li><li>Multiplayer: Two human players take turns</li></ul></li><li>Make Moves: Click on empty squares to place your mark (X or O)</li><li>Win Conditions: Get three marks in a row (horizontal, vertical, or diagonal)</li><li>View Results: Check the scoreboard to see game history and analysis</li></ol><h3>Game Modes</h3>Single Player Mode<ul><li>Player plays as 'X'</li><li>Computer plays as 'O' using an intelligent minimax algorithm</li><li>AI provides challenging gameplay</li></ul>Multiplayer Mode<ul><li>Two players alternate turns</li><li>Player 1 uses 'X', Player 2 uses 'O'</li><li>Perfect for local multiplayer gaming</li></ul><h3>Features Overview</h3>Scoreboard<ul><li>Automatically tracks all game outcomes</li><li>Stores results in scoreboard.csv</li><li>Displays complete game history in a scrollable window</li></ul>Data Analysis<ul><li>Generates pie charts showing win percentages</li><li>Uses pandas for data processing</li><li>Matplotlib for visualization</li><li>Helps players track their performance over time</li></ul>AI AlgorithmThe single-player mode uses the minimax algorithm to provide intelligent gameplay:<ul><li>Evaluates all possible moves</li><li>Chooses the optimal move to maximize winning chances</li><li>Provides challenging but fair gameplay</li></ul><h3>Future improvements:</h3><ul><li>Enhanced AI difficulty levels</li><li>Better UI/UX design</li><li>Additional game statistics</li><li>Sound effects</li><li>Online multiplayer support</li></ul>

Tic Tac Toe Game

<h3>Features</h3><ul><li>Course Management: Create and manage multiple courses</li><li>Attendance Recording: Record attendance for multiple classes per day</li><li>Attendance Tracking: Track attendance status (Present, Absent, Cancelled)</li><li>Progress Monitoring: View attendance percentage for each course</li><li>Class Details: Track time duration and date for each class</li><li>User-Friendly Interface: Clean and intuitive web interface</li></ul><h3>Tech Stack</h3><ul><li>Backend: Django 5.0.7</li><li>Database: SQLite3 (default)</li><li>Frontend: HTML, CSS, Bootstrap (via Django templates)</li><li>Package Management: Pipenv</li><li>Python Version: 3.12</li></ul><h3>Usage</h3><h3>Creating Courses</h3><ol><li>Navigate to the home page</li><li>Click on "Create Course"</li><li>Enter the course name</li><li>Submit to create the course</li></ol><h3>Recording Attendance</h3><ol><li>From the home page, click "Submit Attendance"</li><li>Select the number of classes attended (1-6)</li><li>For each class, fill in:<ul><li>Course name</li><li>Time duration (format: HH:MM-HH:MM, e.g., 14:45-15:45)</li><li>Attendance status (YES/NO/CANCELLED)</li></ul></li><li>Submit the attendance records</li></ol><h3>Viewing Course Details</h3><ol><li>Click on any course from the home page</li><li>View detailed attendance records</li><li>See attendance percentage and statistics</li></ol><h3>Future Enhancements</h3><ul><li>User authentication and authorization</li><li>Export attendance data to CSV/PDF</li><li>Email notifications for low attendance</li><li>Mobile-responsive design improvements</li><li>Attendance analytics and charts</li><li>Bulk import/export functionality</li><li>Calendar integration</li><li>Attendance reminders</li></ul>

AttendanceTracker

<h3>About This Project</h3>This repository, 'HackOpenEnv', is dedicated to the OpenEnv Hackathon. It focuses on building and validating a real-world Reinforcement Learning (RL) environment.<h3>Key Components:</h3><ul><li>rl_env/: This directory contains the main submission environment, specifically designed for the 'Bureaucracy Escape Room' challenge. It includes the inference script, Docker setup for environment execution, and validation scripts to ensure the environment functions correctly.</li></ul><a target="_blank" rel="noopener noreferrer nofollow ugc" class="break-all link-break text-gray-500 hover:underline" href="https://huggingface.co/spaces/Alokik089/bureaucracy_escape_room">Hugging Face Spaces: https://huggingface.co/spaces/Alokik089/bureaucracy_escape_room</a>

OpenEnv Hackathon: Bureaucracy Escape Room

<h3>Key Features</h3><ul><li>Local Inference: Runs STT (Whisper), LLM (Llama 3.2 1B), and TTS (Piper) entirely on your local machine.</li><li>Smart Turn-Taking: Energy-based VAD (Voice Activity Detection) handles natural conversation flow.</li><li>Barge-In Support: The AI stops talking immediately if the caller interrupts.</li><li>Advanced Memory:<ul><li>Ground Truth: Authority context for factual accuracy.</li><li>Profile Memory: Remembers names, facts, and preferences across calls.</li><li>Semantic Context: Retrieves relevant past conversation snippets.</li></ul></li><li>Telephony: Seamless integration with Twilio Voice via Media Streams.</li></ul><h3>Tech Stack</h3><ul><li>Backend: FastAPI (Python 3.10+)</li><li>STT: faster-whisper</li><li>LLM: <a target="_blank" rel="nofollow" class="break-all link-break text-gray-500 hover:underline" href="https://ollama.com">Ollama</a> (Llama 3.2 1B)</li><li>TTS: <a target="_blank" rel="noopener noreferrer nofollow ugc" class="break-all link-break text-gray-500 hover:underline" href="https://github.com/rhasspy/piper">Piper</a></li><li>Audio: NumPy, Scipy</li></ul><h3>Performance expectations (CPU)</h3>Stage LatencyTranscription (Whisper) ~600ms Brain (LLM First Token) ~800ms Voice (TTS Generation) ~150msTotal Response Time~1.5s - 2.5s

AI Voice Calling Assistant

<ul><li>CtrlX is a role-based bug tracking system designed for developers, testers, and project managers to report, assign, and resolve bugs collaboratively. </li><li>The project supports real-time tracking, email notifications, PDF/CSV export, file uploads, and an interactive dashboard. </li><li>Integrated real-time chat via GPT-3.5 API and deployed on Render (backend) &amp; Vercel (frontend). </li><li>The most important feature of the project is its real-time chatting system in custom formed teams integrated with AI to discuss issues and resolve them efficiently.</li></ul>