End-to-end project development for CSE students — real source code, full documentation, synopsis, PPT, and viva support. Every project built from scratch, never recycled.
Full SIEM platform with brute-force detection, privilege escalation monitoring, DNS tunneling detection, file integrity monitoring with SHA-256 baselines, and compliance reporting for PCI DSS, HIPAA & GDPR. Live web dashboard with real-time alerts.
Federated platform for secure threat intelligence sharing across distributed nodes. Features RSA-2048 encryption with digital signatures, automated threat scoring, peer trust scoring, and MITRE ATT&CK Framework integration.
Comprehensive network scanner supporting TCP SYN/FIN/Null/Xmas/ACK, UDP, and SCTP modes. Service/version detection, OS fingerprinting, scripting engine for custom tasks, and stealth evasion techniques via a web-based UI.
Anonymous real-time messaging using AES-256-GCM encryption, PBKDF2-HMAC-SHA256 key derivation (100k iterations), bcrypt password hashing, WebSocket messaging, and encrypted file sharing up to 16MB. Built for local network deployment.
Live network traffic analyzer with web dashboard. Supports TCP, UDP, DNS, HTTP, HTTPS, and ICMP. Detects port scans, tracks suspicious IPs, and uses session-based access with rotating random security keys per session.
Production Python package for generating end-to-end encrypted QR codes with forward secrecy. Each QR uses ephemeral X25519 keys, ChaCha20-Poly1305 / AES-GCM authenticated encryption, and time-based anti-replay TTL protection.
LAN messaging with four-layer encryption: Morse + XOR + AES-256-GCM + HMAC-SHA256. WebSocket-based instant messaging, secure file sharing up to 32MB, and QR code for mobile access. Demonstrates layered security design principles.
First-of-kind browser extension defending against malicious extensions. Monitors installed extensions for suspicious behavior, blocks unauthorized credential theft attempts, delivers real-time threat notifications with minimal performance overhead.
Peer-to-peer file transfer using WebRTC for direct device-to-device connections with no server relay. AES encryption, QR code session sharing, real-time progress tracking, multi-file support up to 16MB each, and HTTPS with SSL/TLS.
Cross-platform secure file deletion tool encrypting files with AES-256 before wiping using multiple overwrite patterns: Random, Zeros, Ones, DoD 5220.22-M, and Gutmann 35-pass. Panic mode, metadata cleanup, and Windows context menu integration.
Zero-database lightweight file server with a terminal-inspired UI. HTTPS authentication, IP-based brute force protection, directory traversal prevention, and real-time file search. Generates QR codes (ASCII + web) for quick access.
Web app for storing time-locked secrets with Fernet encryption plus optional passphrase double-lock. Supports QR code sharing, real-time countdown timers, file uploads up to 10MB, and automatic one-time access cleanup.
Terminal-styled Firefox extension for advanced cookie management. Full CRUD operations, advanced search and filtering, bulk operations, JSON import/export, and visual security indicators for Secure, HttpOnly, and SameSite flags.
Firefox extension detecting clipboard hijacking attacks via real-time SHA-256 hash monitoring. Uses Levenshtein distance analysis to eliminate false positives, and provides configurable sensitivity visual alerts.
Monitors all outgoing browser network requests in real time and assigns a 0–100 privacy risk score. Stores request history in IndexedDB with automatic cleanup, supports JSON export, and features a developer-focused monospace UI.
Automates proxy switching via proxychains4 for IP rotation with geo-IP lookup, proxy latency testing, free proxy scraping from multiple public sources, and configurable rotation intervals with network-change triggers.
Browser-based tool to hide AES-256-GCM encrypted messages inside PNG images using LSB and LSB Matching algorithms. Fully client-side with zero server communication, PBKDF2 key derivation (100k iterations), and drag-and-drop support.
Anti-fingerprinting extension generating unique stable profiles per website. Injects Canvas/WebGL noise, spoofs navigator and screen properties, defends audio fingerprinting, blocks WebRTC leaks, and offers Balanced/Paranoid/Compatibility modes.
Windows battery management utility with profile-based power control. Unlimited custom profiles, automation triggers on battery level/app launch/network change/time, Windows services management, and real-time battery analytics.
GUI application providing anti-AFK functionality with stealth capabilities. Features 6 tabs (Main, Advanced, Stealth, Session, Monitor, Presets), multiple activation modes, human-like pattern simulation, and global hotkey support.
ML pipeline trained on PE file static features and behavioral datasets to classify binaries as benign or malicious. FastAPI endpoint for file upload, real-time prediction with confidence scores, SHAP explainability reports, and analytics dashboard.
BERT-based NLP classifier for phishing email detection. Supports bulk CSV upload, real-time email header analysis, per-email risk scoring, and an admin dashboard with historical trend charts and model evaluation metrics.
ML-based IDS trained on NSL-KDD and CICIDS2017 datasets detecting DoS, Probe, R2L, and U2R attacks using a Random Forest + LSTM hybrid model. Integrates live Scapy packet capture with real-time alert notifications.
CNN + transfer learning model trained on authentic and DeepFake facial datasets to detect subtle manipulation artifacts. Image and video frame analysis, confidence scoring, batch upload support, and results dashboard with heatmap overlays.
Predicts risk for multiple conditions (diabetes, heart disease, Parkinson's) from patient symptom and lab input. Separate trained classifiers per disease, unified Flask interface, confidence score output, and health history tracking dashboard.
Retrieval-Augmented Generation chatbot allowing users to upload documents (PDF, DOCX) and ask questions in natural language. Vector embeddings for semantic search, LangChain orchestration, and a clean chat UI with source citation.
Real-time object detection using YOLOv8 supporting webcam feed, uploaded images, and video files. Detects 80+ COCO classes with bounding boxes and confidence scores. Includes web interface, custom class filtering, and detection history export.
Fraud detection pipeline on imbalanced transaction datasets using SMOTE for class balancing and ensemble models (Random Forest, XGBoost, Isolation Forest). Real-time API for transaction scoring, threshold tuning, and analyst dashboard.
Parses uploaded resumes, extracts skills and experience using NLP, and ranks them against a job description using cosine similarity. Suggests relevant job roles, highlights skill gaps, and generates a match percentage report per candidate.
AI-powered study assistant that ingests course PDFs, generates topic summaries, creates MCQ quizzes, and answers student queries via RAG. Agent workflow auto-generates flashcards and provides citation-backed explanations.
NLP pipeline using TF-IDF vectorization with Naive Bayes and SVM models to classify SMS as spam or legitimate. Training scripts, evaluation metrics (precision, recall, F1, ROC), pickle model export, and simple CLI inference tool.
Automates attendance by detecting and recognizing faces in real time via webcam. Multi-person enrollment from photos, handles multiple faces per frame, logs attendance to CSV, and provides a Tkinter management dashboard.
Regression and classification model predicting final student grades from attendance, internal marks, and activity data. Flask form for single predictions, model comparison charts, and CSV batch prediction mode.
Analyzes sentiment of submitted text or scraped social media content using VADER and a fine-tuned DistilBERT model. Displays real-time sentiment trend charts, word clouds, and emotion classification breakdowns on a clean web dashboard.
Predicts residential property prices from location, size, and amenity features using Linear Regression, Ridge, and XGBoost. Interactive Flask UI for custom inputs, feature importance plots, and model performance comparison charts.
Recommends movies based on user preferences using content-based (TF-IDF on metadata) and collaborative filtering (SVD matrix factorization). Streamlit UI with genre filtering, rating input, and top-N recommendations display.
Classifies news articles as real or fake using TF-IDF and PassiveAggressiveClassifier / Logistic Regression. Flask web interface for URL or paste input, confidence score display, and a history log of analyzed articles.
CNN model trained on the MNIST dataset to recognize handwritten digits with 99%+ test accuracy. Tkinter canvas where users draw digits in real time and get instant classification with confidence bars.
Predicts whether a telecom or SaaS customer will churn using Logistic Regression, Random Forest, and XGBoost trained on the Telco dataset. Feature importance visualization, probability scores, and Flask dashboard for business teams.
LSTM recurrent neural network model for forecasting stock closing prices using historical OHLCV data from Yahoo Finance. Multi-step prediction, train/test split visualization, confidence intervals, and Flask web UI with interactive Plotly charts.