Projects
| Academic Year | Group Number | Title of the Project | Problem statement / Complex Engineering Problem | Student Name | Guide Name | In Association with Industry (If Yes) Industry Name & Industry Supervisor Name | Relevant SDGs | PO Mapping | PSO Mapping | Project Report ( Turnitin Ai Checker) | Project Report ( Turnitin Similarity Checker) | Achievement |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| AY 2025-26 | AIDS-1 | Engunity X : AI-Powered SaaS Platform for Research, Code Generation & Data Analysis | Engunity X aims to solve the problem of fragmented AI tools by providing a unified platform for research assistance, code generation, and data analysis. The system integrates cloud-based LLMs with local fallback models to ensure reliable, low-latency, and cost-efficient AI services. It incorporates Retrieval-Augmented Generation (RAG) using FAISS for intelligent document-based question answering and Docker sandboxing for secure code execution. The platform also addresses security and transparency challenges through blockchain-based features such as decentralized identity (DID), smart contract auditing, and on-chain verification of AI-generated outputs. Developing such a platform involves complex engineering challenges including hybrid AI orchestration, scalable cloud architecture, cybersecurity, distributed systems integration, and secure API management while maintaining scalability, reliability, and data integrity. | Shubh Shah Nigranth Shah Anurag Kumar Shreyash Mokani | Dr.Rahul Pchade | ----- | SDG 9 (Industry, Innovation, and Infrastructure) | PO1, PO2, PO3, PO5, PO10, PO11 | PSO1, PSO2 | View | View | |
| AY 2025-26 | AIDS-2 | AI -Powered Web Scraper Using Large Language Models | The project addresses the limitations of traditional web scraping systems that rely on fragile rule-based techniques such as XPath and CSS selectors, which fail when webpage structures change. Existing approaches also lack contextual understanding and struggle to extract meaningful information from dynamic and unstructured web content. To overcome these challenges, the proposed system integrates Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and multimodal data representation for intelligent web data extraction. The system processes webpages using textual content, DOM structure, and visual layout to generate unified embeddings stored in a FAISS-based retrieval system for context-aware extraction. A cache-aware mechanism is implemented to detect structural similarities between webpages and reuse embeddings, improving efficiency and reducing computational overhead. Developing this system involves complex engineering challenges including multimodal data processing, semantic retrieval optimization, scalable vector indexing, AI-driven information extraction, dynamic webpage handling, and low-latency distributed system design while maintaining scalability, adaptability, and extraction accuracy. | Dev Raval Aryan Rana Aditya Magar Rishith Poojary | Dr.Pravin Shinde | Mr. Ayon Mondal Deepcytes UK-based cyber intelligence company" | SDG 9 (Industry, Innovation, and Infrastructure), SDG 16 (Peace, Justice, and Strong Institutions) | PO2, PO3, PO5, PO8, PO10 | PSO2 | View | View | |
| AY 2025-26 | AIDS-3 | Wallify â Make Walls Talk, Make Vibes Pop. | The Wallify project addresses the challenge of visualizing interior wall decorations before actual implementation, which often leads to poor design decisions and increased cost. Traditional interior planning methods lack interactive visualization and personalized recommendation capabilities, making it difficult for users to select suitable decor arrangements for their spaces. To overcome these limitations, the system integrates Artificial Intelligence (AI), Computer Vision, and Augmented Reality (AR) to provide a âtry-before-you-decorateâ experience. The platform detects wall regions from uploaded images, analyzes surrounding objects, and recommends suitable decorative elements based on user preferences such as theme, event type, and color palette. Using AR-based visualization, users can interactively place, resize, and customize virtual decor objects on real wall images. Developing this system involves complex engineering challenges including image segmentation, object detection, AR rendering, real-time interaction handling, AI-based recommendation systems, and seamless integration of computer vision with immersive visualization technologies while ens | Moksh Gala Vishvesh Ghadge Saniya Shah Akshi Gori | Dr.Pravin Shinde | ----- | SDG 11 (Sustainable Cities and Communities), SDG 12 (Responsible Consumption and Production) | PO3, PO5, PO7, PO9, PO10 | PSO2, PSO3 | View | View | |
| AY 2025-26 | AIDS-4 | Data Fusion based Two stage Cascade Framework for Multi Modality Face Anti Spoofing | Traditional face recognition systems are vulnerable to spoofing attacks such as printed photos, replay videos, and 3D masks, leading to security threats and unauthorized access. Existing anti-spoofing methods often rely on single-modality data and fail under complex real-world conditions. Therefore, a robust multi-modality face anti-spoofing system is required for accurate and secure real-time authentication. | Hamiz Abdul Pathan Harsh Milind Mohite Harish Deepak Shinde Pratham Baraskar | Ms. Bhagyashree Patil | ----- | SDG 9 (Industry, Innovation, and Infrastructure), SDG 16 (Peace, Justice, and Strong Institutions) | PO2, PO4, PO5, PO8 | PSO1, PSO2 | View | View | IEEE Paper Publication |
| AY 2025-26 | AIDS-5 | MCP-Multi-Tenant Chat Support System with AI Integration | The MCP (Multi-Tenant Chat Support System with AI Integration) project addresses the limitations of traditional customer support systems such as delayed responses, poor contextual understanding, limited scalability, and inaccurate AI-generated outputs. The system implements a Retrieval-Augmented Generation (RAG) framework that combines semantic information retrieval with AI-based response generation to provide accurate, context-aware, and citation-backed responses. It supports a secure multi-tenant architecture that enables multiple organizations to operate independently on a shared platform while maintaining strict data isolation. The platform integrates Google Gemini, vector databases, WebSocket-based real-time communication, and cloud-native technologies to ensure scalability, reliability, and low response latency. Developing the system involves complex engineering challenges including real-time distributed communication, AI integration, semantic search optimization, secure authentication, scalable cloud deployment, and efficient resource management while minimizing hallucination and ensuring enterprise-grade security and performance. | Chirag Patankar Preksha Dewoolkar Samriddhi Pande | Dr.Rahul Pchade | ----- | SDG 3 (Good Health and Well-being), SDG 11 (Sustainable Cities and Communities) | PO3, PO6, PO7, PO10 | PSO3 | View | View | |
| AY 2025-26 | AIDS-6 | ResuMind AI-Where resumes meet intelligence | Current AI recruiting systems rely on keyword-based ATS matching, causing qualified candidates to be overlooked. Existing tools are either expensive or lack analytical capability, while interview preparation remains disconnected from resume evaluation. An integrated AI platform is needed for semantic resume scoring, ATS optimization, resume enhancement, CV generation, and personalized interview preparation. | Dhruvi Chheda Bhoomi Chotaliya Aaryan Kamerkar Swayam Karapurkar | Dr.Pravin Shinde | Mr. Manav Rupani Deepcytes UK-based cyber intelligence company" | SDG 8 (Decent Work and Economic Growth), SDG 4 (Quality Education) | PO2, PO3, PO5, PO9, PO10 | PSO2, PSO3 | View | View | IEEE Paper Publication |
| AY 2025-26 | AIDS-7 | MAITRI: An Al Assistant for Psychological & Physical Well-Being of Astronauts | The Maitri AI project addresses the psychological challenges faced by individuals in isolated environments such as space missions, research stations, and long-term healthcare facilities, where limited human interaction can lead to stress, anxiety, loneliness, and emotional distress. Traditional communication systems fail to provide continuous emotional companionship and personalized support in such conditions. To overcome these limitations, the system integrates Artificial Intelligence technologies including Automatic Speech Recognition (ASR), Emotion Recognition, Retrieval-Augmented Generation (RAG), voice cloning, and Large Language Models (LLMs) to create emotionally aware and context-driven conversations. The platform analyzes the userâs emotional state, retrieves relevant conversational memories, and generates empathetic responses delivered through natural-sounding synthesized voices. Developing this system involves complex engineering challenges such as real-time speech processing, emotion detection, multimodal AI integration, memory-based contextual retrieval, low-latency voice synthesis, and resource-efficient deployment while ensuring reliability, personalization, emotional accuracy, and natural human-computer interaction. | Ramchandra Bharat Mulik Ajay Selvan Nadar Tejaswi Yadav Purvesh Gaikwad | Dr.Pravin Shinde | ----- | SDG 3 (Good Health and Well-being) | PO3, PO4, PO6, PO7, PO9 | PSO2, PSO3 | View | View | IEEE Paper Publication |
| AY 2025-26 | AIDS-8 | Chest X-Ray Report Summarization using Vision -Text Transformer | Radiologists spend significant time analyzing chest X-ray images and preparing accurate diagnostic reports, which can lead to delays in clinical decision-making, especially when handling large volumes of scans. Manual report generation requires expert interpretation and is prone to variability. Developing an automated system that can accurately detect thoracic abnormalities from chest X-rays and generate clinically meaningful textual summaries is a complex engineering challenge. The system must effectively integrate computer vision and natural language processing techniques to bridge the gap between image understanding and medical report generation. Additionally, ensuring diagnostic accuracy, clinical relevance, and reliability of generated reports remains a critical challenge in healthcare AI applications. | Aniket Nikalje Vipul Parmar Anshul Pal Kinjay Solanki | Mr. Milind Khairnar | ----- | SDG 3 (Good Health and Well-being) | PO1, PO3, PO4, PO5, PO8 | PSO1, PSO2, PSO3 | View | View | IEEE Paper Publication |
| AY 2025-26 | AIDS-9 | We Care : Smart Healthcare AI | The WeCare (Smart Healthcare AI) project addresses the increasing mental health issues among students and the limitations of traditional mental health assessment methods, which are often time-consuming, less accessible, and may increase social stigma. Early detection and continuous monitoring of emotional well-being remain major challenges in educational environments. To overcome these issues, the system proposes an AI-powered chatbot platform that evaluates student mental health through conversational analysis, facial emotion recognition, and structured psychological questionnaires. The platform analyzes user interactions and emotional patterns to generate assessment reports along with expert-based recommendations and self-awareness insights. Developing this system involves complex engineering challenges including emotion detection, natural language processing, AI-driven behavioral analysis, facial expression recognition, secure handling of sensitive data, ethical AI implementation, and real-time assessment generation while ensuring accuracy, privacy, reliability, and responsible use of AI in mental healthcare. | Soham Sandeep Bagdane Sahil Bharat Dharne Vrushabh Kailas Dhomse Harsh Deepak Parmar | Ms. Bhagyashree Patil | ----- | SDG 3 (Good Health and Well-being), SDG 11 (Sustainable Cities and Communities) | PO3, PO6, PO7, PO9, PO10 | PSO3 | View | View | IEEE Paper Publication |
| AY 2025-26 | AIDS-10 | AI mock Interview | The AI-powered mock interview system addresses the limitations of traditional interview preparation platforms that rely on static and generic question banks without considering a candidateâs background and skillset. The system analyses uploaded resumes using a DeBERTa-based NER model to extract structured information such as skills, projects, and job roles, and generates personalized interview questions using the Mistral LLM. It supports real-time voice-based interaction through WebRTC, Whisper speech recognition, voice activity detection, and Coqui TTS for natural conversational interviews. The platform further evaluates candidate responses using a five-dimensional feedback framework that measures relevance, technical accuracy, depth, clarity, and confidence. Developing this system involves complex engineering challenges including NLP-based resume parsing, low-latency real-time audio processing, AI-driven question generation, multimodal feedback analysis, speech processing, and scalable AI system integration while ensuring accuracy, responsiveness, and reliable performance. | Drashti Gajra Krisha Bhanushali Tushar Patel Samiksha Thakur | Dr.Rahul Pchade | Mr. Manav Rupani Deepcytes UK-based cyber intelligence company" | SDG 4 (Quality Education) | PO1, PO2, PO3, PO5, PO10 | PSO1, PSO2 | View | View | |
| AY 2025-26 | AIDS-11 | SAGE: First aid Chatbot | The SAGE-First Aid Bot project addresses the challenge of providing immediate and accurate first aid guidance during medical emergencies where people often panic or lack proper medical knowledge. Delayed or incorrect first aid responses can increase the severity of injuries and risk to human life, especially when professional medical assistance is not immediately available. To overcome these limitations, the system uses Artificial Intelligence and Natural Language Processing (NLP) to deliver conversational and context-aware first aid guidance based on user-reported symptoms or emergency situations. The chatbot provides appropriate safety measures, emergency procedures, and first aid instructions for conditions such as burns, bleeding, fractures, choking, and heart attacks. Developing this system involves complex engineering challenges including real-time NLP-based interaction, medical knowledge representation, emergency response decision-making, conversational AI design, scalable deployment, and ensuring reliability and accuracy of medical recommendations while maintaining accessibility, usability, and public safety. | Vinit Sangoi Sachi Pandya Het Shah Harsh Shinde | Dr. Manimala Mahato | ----- | SDG 3 (Good Health and Well-being) | PO3, PO5, PO6, PO9, PO10 | PSO2, PSO3 | View | View | |
| AY 2025-26 | AIDS-12 | Lunar Boulder Detection | Develop a novel algorithm to detect and analyze individual boulders on the lunar surface using Chandrayaan-1 and Chandrayaan-2 imagery. The algorithm should improve upon existing methods by accurately identifying boulders, estimating their size and spatial distribution, and distinguishing them from background terrain. It should also attempt to trace boulders back to likely sources (e.g., craters or cliffs) and highlight regions with recent or active boulder falls. The goal is to generate new insights into lunar boulder distribution, origins, and surface activity. | Dhruv Solanki Shrawani Sawant Shashank Singh Sudhanshu Singh | Ms. Swati Uparkar | ----- | SDG 9 (Industry, Innovation, and Infrastructure) | PO2, PO3, PO4, PO5, PO7 | PSO1, PSO2, PSO3 | View | ||
| AY 2025-26 | AIDS-13 | Nova Ai : Personalized Course Generator | Designing an AI-powered platform that generates personalized learning roadmaps and course recommendations based on user interests, skills, goals, and learning pace. The system addresses challenges such as adaptive content generation, recommendation accuracy, scalable AI integration, user engagement, and intelligent curriculum structuring using LLMs and machine learning techniques. | Vivek Pundalik Akate Sheedhukumar Sudarshan Achary ANSHUL ABHIMANU CHAUHAN Prem Yogesh Dasani | Dr.Pravin Shinde | Mr. Manav Rupani Deepcytes UK-based cyber intelligence company" | SDG 4 â Quality Education, SDG 9 â Industry, Innovation and Infrastructure | PO2, PO3, PO5, PO10 | PSO2, PSO3 | View | View | IEEE Paper Publication |
| AY 2025-26 | AIDS-14 | FitVision: Vision -Based Intelligent Fitness Assistance Using Pose Guided Movement Analysis | Designing an AI-powered fitness monitoring system that uses computer vision and pose estimation to track body movements, analyze exercise posture, and count repetitions in real time. The system addresses challenges such as posture detection accuracy, movement analysis under varying body structures and camera conditions, real-time processing, scalable AI integration, and user interaction through a domain-specific fitness chatbot using machine learning and LLM-based technologies. | Aashna anchan Purva Ambre Hasan contractor Husain contractor | Ms. Swati Uparkar | ----- | SDG 3 (Good Health and Well-being) | PO3, PO5, PO6, PO7 | PSO2, PSO3 | View | ||
| AY 2025-26 | AIDS-15 | Auto ITR | Filing Income Tax Returns (ITR) is a complicated and time-consuming process for many individuals due to manual calculations, lack of tax knowledge, and chances of errors. Users often struggle with understanding tax rules, managing documents, and identifying deductions. AutoITR aims to simplify the tax filing process by providing an automated, user-friendly platform that calculates taxes, validates data, and assists users in filing accurate ITRs quickly and efficiently. | Haireet Mehta Shubham Jain Shreyansh Jain Kaushal Mhatre | Mr. Milind Khairnar | ----- | SDG 8 (Decent Work and Economic Growth), SDG 16 (Peace, Justice, and Strong Institutions) | PO1, PO2, PO3, PO5, PO11 | PSO2 | View | View |
UG Program in Electronics and Telecommunication Engineering is accredited by NBA for 3 years from Academic Year 2024-25, 2025-26, and 2026-27, i.e., up to 30/06/2027.
UG Program in Computer Engineering and Information Technology is accredited by NBA for 3 years from Academic Year 2025-26, 2026-27, and 2027-28, i.e., up to 30/06/2028.
