TechFrontier Problem Statements

• A twin bot AI assistant students can consult when their mentor is unavailable. • Students ask questions; bot answers coherently and can generate additional explanations or examples. • Option to generate a PDF document that summarizes answers / explanations for deeper understanding. • Tracks the topics the student has covered to help tailor responses.

Problem: Many students struggle with generic courses that don't match their skills. Challenge: Build an AI system that analyzes a student's past performance, interests, and learning style to generate a personalized learning path with course recommendations, quizzes, study schedules and provides instant explanations, examples, or hints. Deliverables: • Adaptive learning recommendations • AI-generated quiz questions • Progress tracking dashboard • Context-aware doubt resolution • Ability to handle multiple subjects/topics • Integration with text, video, and image-based content

Theme: Network Access Security Problem Statement: Implement a Zero Trust Network (ZTN) framework that enforces continuous authentication and authorization using standards like OAuth2.0, OpenID Connect, and FIDO2. Requirements: • Continuous verification of user identity, device health, and context • Implement micro-segmentation to control internal access • Support adaptive access control (location, device, behavior) • Log all access attempts for auditing • Provide centralized policy engine for access control Bonus Challenges: • Integrate AI-driven risk scoring for adaptive policy decisions • Support hardware-based authentication (YubiKey/FIDO2) • Simulate an enterprise network environment for testing

Theme: Post-Incident Investigation Problem Statement: Build a digital forensics tool to extract, analyze, and recover evidence from compromised systems. The system should support metadata extraction, deleted file recovery, and log correlation. Requirements: • Extract metadata (timestamps, file owners, access info) • Recover deleted or partially corrupted files • Analyze system logs for malicious activities • Generate investigation timeline automatically • Export evidence reports for legal documentation Bonus Challenges: • Implement hash comparison for file integrity verification • Integrate with Volatility for memory forensics • Support multi-disk or image-based investigations

Theme: Logistics & Traceability Problem Statement: Create a blockchain-powered system for tracking products from manufacturing to delivery, ensuring transparency and authenticity in the supply chain. Requirements: • Each product batch recorded as a blockchain transaction • Role-based access (supplier, distributor, retailer, consumer) • QR code tracking for every stage • Immutable logs of movement, timestamp, and handlers • Real-time dashboard for product traceability Bonus Challenges: • Integrate IoT sensors for automatic data updates • Use smart contracts for automated payments • Visualize supply chain flow with 3D/AR interface

Theme: Web Application Security Problem Statement: Build an automated vulnerability scanner that tests web applications for the OWASP Top 10 vulnerabilities and generates detailed reports with fix recommendations. Requirements: • Scan for SQL injection, XSS, CSRF, insecure deserialization, etc. • Support both authenticated and unauthenticated scans • Generate detailed vulnerability reports with severity and fix suggestions • Implement scheduling for periodic scans • Provide dashboard with vulnerability statistics per application Bonus Challenges: • Integrate with DevSecOps pipelines • Implement REST/GraphQL API scanning • Support automated retesting after fixes

Theme: Mobile Data Security Problem Statement: Create a secure file vault app that encrypts user files using Android Keystore, ensuring data protection even if the device is rooted or stolen. Requirements: • Encrypt/decrypt files using Keystore keys • Biometric (fingerprint/face) authentication for access • Prevent screenshots and background recording • Secure file deletion (overwrite and remove traces) • Display file access logs Bonus Challenges: • Integrate Google Drive/Dropbox backup with encrypted sync • Support decoy password for stealth mode • Detect and alert on root/jailbreak attempts

Background: Deepfakes are used in fraudulent banking calls, KYC bypass, and misinformation campaigns. Current detection is mostly offline. Challenge: Build a real-time multimodal deepfake detector (voice + video + behavior analysis). Expected Outcome: Verification layer for live video calls or financial KYC workflows. Datasets: • Face Forensics++ – https://github.com/ondyari/FaceForensics • ASVspoof 2019 – https://www.asvspoof.org/ • Deep Fake Detection Challenge (DFDC) – https://ai.meta.com/datasets/dfdc/

Theme: Mobile Data Security Problem Statement: Create a secure file vault app that encrypts user files using Android Keystore, ensuring data protection even if the device is rooted or stolen. Requirements: • Encrypt/decrypt files using Keystore keys • Biometric (fingerprint/face) authentication for access • Prevent screenshots and background recording • Secure file deletion (overwrite and remove traces) • Display file access logs Bonus Challenges: • Integrate Google Drive/Dropbox backup with encrypted sync • Support decoy password for stealth mode • Detect and alert on root/jailbreak attempts

Theme: Authentication Security Problem Statement: Create a detection and prevention system for credential stuffing attacks using behavioral analytics and anomaly detection. Requirements: • Detect repeated login attempts with known credential patterns • Implement IP/device fingerprinting • Integrate CAPTCHA or MFA after suspicious activity • Provide real-time alerts and blocking • Maintain audit trail of detected incidents Bonus Challenges: • Train ML models using known credential stuffing datasets • Integrate with threat intelligence feeds • Visualize login anomalies on a heatmap dashboard

Theme: Endpoint Security & Response Problem Statement: Design an intelligent ransomware detection system that can identify abnormal file encryption activity and automatically isolate infected endpoints to prevent network-wide spread. Requirements: • Monitor filesystem activity for encryption patterns • Detect abnormal CPU/disk usage and unauthorized file access • Implement automated containment (e.g., network isolation or process kill) • Use ML anomaly detection for real-time ransomware behavior recognition • Provide centralized dashboard for alerts and containment logs Bonus Challenges: • Implement sandbox analysis for unknown executables • Integrate with SIEM systems (e.g., Splunk/ELK) • Simulate ransomware attacks safely for testing detection accuracy

• Face recognition-based attendance: employees' faces are recognized in a specific location to mark presence automatically. • Track number of days present, total working hours per day/week/month. • Leave application & approval workflow. • Pay slip downloads. • Announcements, document request system, calendar of events/holidays, recent activity logs.

• Product inventory management: listing products, tracking stock quantities, adjusting stock after each sale. • Billing / invoice generation when products sold; apply codes, taxes, discounts. • Track product sales over time; generate reports (daily, weekly, monthly). • Predictive insights / suggestions to increase sales: slow-moving products, bundling, special offers. • System to send special offers/promotions to customers.

• A public facing website for the business/agency, showcasing services (e.g. SaaS, web design, graphic design, marketing). • Built with only HTML, CSS, JavaScript, PHP (no external JS/CSS animation libraries). • Animations hand-coded. • Ability for business clients to upload their brand logo + type of business, and in response the system generates a basic website template or poster suited to them. • Contact form that can send messages to their WhatsApp. • Admin dashboard to collect user leads/data, track which pages/services are viewed, track traffic. • Portfolio page that is clean, well laid out, responsive design.

Build a solution that applies unsupervised machine learning (e.g., clustering, autoencoders) to detect anomalies in encrypted network traffic metadata (packet sizes, timing, frequency) without decrypting the payload, helping to spot malicious behaviors like data exfiltration or command-and-control (C2) communication. Objective: Apply AI/ML to detect threats in encrypted environments where traditional signature-based detection fails. Input: Encrypted network traffic (PCAP files), connection metadata. Output: Anomaly score, flagged suspicious sessions, visualization of traffic patterns. Bonus: Self-learning model that improves detection accuracy over time.

• Product inventory management: listing products, tracking stock quantities, adjusting stock after each sale. • Billing / invoice generation when products sold; apply codes, taxes, discounts. • Track product sales over time; generate reports (daily, weekly, monthly). • Predictive insights / suggestions to increase sales: slow-moving products, bundling, special offers. • System to send special offers/promotions to customers.

• A twin bot AI assistant students can consult when their mentor is unavailable. • Students ask questions; bot answers coherently and can generate additional explanations or examples. • Option to generate a PDF document that summarizes answers / explanations for deeper understanding. • Tracks the topics the student has covered to help tailor responses.

• A twin bot AI assistant students can consult when their mentor is unavailable. • Students ask questions; bot answers coherently and can generate additional explanations or examples. • Option to generate a PDF document that summarizes answers / explanations for deeper understanding. • Tracks the topics the student has covered to help tailor responses.

A trustless marketplace for freelancers and clients where smart contracts hold payments and release them based on milestone completion or dispute resolution through decentralized arbitration. • No platform fees (unlike Upwork/Fiverr taking 20%) • Reputation system stored on-chain • Automated payments upon work delivery • Community-based dispute resolution

Build a solution that applies unsupervised machine learning (e.g., clustering, autoencoders) to detect anomalies in encrypted network traffic metadata (packet sizes, timing, frequency) without decrypting the payload, helping to spot malicious behaviors like data exfiltration or command-and-control (C2) communication. Objective: Apply AI/ML to detect threats in encrypted environments where traditional signature-based detection fails. Input: Encrypted network traffic (PCAP files), connection metadata. Output: Anomaly score, flagged suspicious sessions, visualization of traffic patterns. Bonus: Self-learning model that improves detection accuracy over time.

Theme: Data Authenticity Verification Problem Statement: Develop a blockchain-based system that verifies the integrity of files by storing their hash values immutably on a distributed ledger. Requirements: • Generate SHA-256 or SHA-3 hash for uploaded files • Record hash on blockchain (Ethereum/Polygon) • Verify file authenticity by comparing hash values • Display transaction logs and verification results • Provide API endpoints for integration Bonus Challenges: • Support batch verification and scheduled checks • Integrate IPFS for decentralized file storage • Implement tamper alert notifications

• A twin bot AI assistant students can consult when their mentor is unavailable. • Students ask questions; bot answers coherently and can generate additional explanations or examples. • Option to generate a PDF document that summarizes answers / explanations for deeper understanding. • Tracks the topics the student has covered to help tailor responses.

Problem Statement: Design and implement a secure banking API system that handles real-time fund transfers, account balance checks, and transaction history. The system must implement JWT-based authentication with access tokens (15-minute expiry) and refresh tokens (7-day expiry). Requirements: • Implement user login endpoint that returns both access and refresh tokens • Create protected endpoints for: balance inquiry, fund transfer, transaction history, and beneficiary management • Implement token refresh mechanism without requiring re-login • Handle concurrent transaction requests to prevent race conditions • Implement rate limiting (max 100 requests per minute per user) • Add real-time transaction notifications using WebSockets • Ensure atomic transactions (transfer should debit one account and credit another as a single operation) • Implement token blacklisting for logout functionality • Add 2FA verification for high-value transactions (>$10,000) Bonus Challenges: • Implement token rotation (issue new refresh token on each refresh) • Add device fingerprinting to bind tokens to specific devices • Implement suspicious activity detection (multiple failed auth attempts)

• A public facing website for the business/agency, showcasing services (e.g. SaaS, web design, graphic design, marketing). • Built with only HTML, CSS, JavaScript, PHP (no external JS/CSS animation libraries). • Animations hand-coded. • Ability for business clients to upload their brand logo + type of business, and in response the system generates a basic website template or poster suited to them. • Contact form that can send messages to their WhatsApp. • Admin dashboard to collect user leads/data, track which pages/services are viewed, track traffic. • Portfolio page that is clean, well laid out, responsive design.

Theme: Intrusion Deception Systems Problem Statement: Develop a honeypot system that detects and logs unauthorized access attempts, analyzes attacker behavior, and sends alerts to administrators. Requirements: • Deploy multiple honeypots (SSH, HTTP, FTP, etc.) • Log attacker commands, IP, and timestamps • Analyze attack patterns and generate behavior profiles • Alert administrators in real time via dashboard or email • Implement visualization for attack trends Bonus Challenges: • Integrate ML for attack classification (bot, script, human) • Generate fake data environments to mislead attackers • Correlate honeypot data with firewall logs

Theme: Intelligent Network Monitoring Problem Statement: Develop an AI-powered system that monitors real-time network traffic and detects anomalies, potential intrusions, or data exfiltration attempts. Requirements: • Capture live network packets and extract features • Use ML algorithms for anomaly detection • Classify normal vs malicious traffic • Provide dashboards for visualization (traffic flow, protocol stats) • Support alerts on detected anomalies Bonus Challenges: • Integrate with SNMP or NetFlow for enterprise monitoring • Implement explainable AI to justify anomaly decisions • Simulate DDoS or port scan attacks for testing

Problem & Challenge: Children in crowded areas risk getting lost. Develop an IoT system that tracks children in real time, detects when they leave safe zones, and notifies guardians instantly. The system should be accurate, low-power, and reliable. Objectives & Scope: • Wearable devices using GPS, RFID, BLE/LoRa for real-time tracking • Geofencing alerts for safe-zone breaches • Mobile/web platform for monitoring and notifications • Optimize low power consumption and validate through simulation and real-world testing

Theme: Augmented Reality Navigation Problem Statement: Build an Android AR app that helps users navigate both indoor and outdoor environments using real-world overlays and 3D path guidance. Requirements: • Use ARCore for 3D navigation overlays • Integrate GPS for outdoor navigation • Use Wi-Fi or Bluetooth beacons for indoor positioning • Display arrows and labels in AR view • Include search for specific rooms, buildings, or points of interest Bonus Challenges: • Add voice-assisted navigation • Enable multi-floor navigation • Support offline maps and routes caching

• Product inventory management: listing products, tracking stock quantities, adjusting stock after each sale. • Billing / invoice generation when products sold; apply codes, taxes, discounts. • Track product sales over time; generate reports (daily, weekly, monthly). • Predictive insights / suggestions to increase sales: slow-moving products, bundling, special offers. • System to send special offers/promotions to customers.

Problem: Many students struggle with generic courses that don't match their skills. Challenge: Build an AI system that analyzes a student's past performance, interests, and learning style to generate a personalized learning path with course recommendations, quizzes, study schedules and provides instant explanations, examples, or hints. Deliverables: • Adaptive learning recommendations • AI-generated quiz questions • Progress tracking dashboard • Context-aware doubt resolution • Ability to handle multiple subjects/topics • Integration with text, video, and image-based content

Theme: Data Confidentiality & Access Control Problem Statement: Develop an end-to-end encrypted file sharing system that ensures only authorized recipients can view or download shared files, even if the storage server is compromised. Requirements: • Implement client-side encryption using AES/RSA • User authentication and access control (link expiry, revoke access) • Store files securely with metadata encryption • Support sharing via temporary access tokens • Maintain activity logs for audit purposes Bonus Challenges: • Integrate blockchain for immutable access logs • Support large file uploads with chunked encryption • Add watermarking and access tracking

Theme: Data Confidentiality & Access Control Problem Statement: Develop an end-to-end encrypted file sharing system that ensures only authorized recipients can view or download shared files, even if the storage server is compromised. Requirements: • Implement client-side encryption using AES/RSA • User authentication and access control (link expiry, revoke access) • Store files securely with metadata encryption • Support sharing via temporary access tokens • Maintain activity logs for audit purposes Bonus Challenges: • Integrate blockchain for immutable access logs • Support large file uploads with chunked encryption • Add watermarking and access tracking

Anti-counterfeiting system for high-end products (watches, handbags, sneakers) with verifiable authenticity and ownership history. • Each item gets a unique blockchain identity at manufacture • Track entire ownership chain (useful for resale) • Prove authenticity instantly via QR code • Combat fake luxury goods market • Warranty and repair history tracking

Theme: Cyber Awareness & Training Problem Statement: Design an Android app that simulates real-world cybersecurity attack scenarios to train users in identifying phishing, malware, and social engineering threats. Requirements: • Interactive phishing email and message simulations • Score-based evaluation for each simulation • Educational feedback after each attempt • Admin panel for generating new training scenarios • Track user progress and awareness improvement Bonus Challenges: • Integrate AR for physical security awareness scenarios • Implement gamified leaderboard system • Generate phishing templates using AI

• Product inventory management: listing products, tracking stock quantities, adjusting stock after each sale. • Billing / invoice generation when products sold; apply codes, taxes, discounts. • Track product sales over time; generate reports (daily, weekly, monthly). • Predictive insights / suggestions to increase sales: slow-moving products, bundling, special offers. • System to send special offers/promotions to customers.

• Face recognition-based attendance: employees' faces are recognized in a specific location to mark presence automatically. • Track number of days present, total working hours per day/week/month. • Leave application & approval workflow. • Pay slip downloads. • Announcements, document request system, calendar of events/holidays, recent activity logs.

• Face recognition-based attendance: employees' faces are recognized in a specific location to mark presence automatically. • Track number of days present, total working hours per day/week/month. • Leave application & approval workflow. • Pay slip downloads. • Announcements, document request system, calendar of events/holidays, recent activity logs.

• A twin bot AI assistant students can consult when their mentor is unavailable. • Students ask questions; bot answers coherently and can generate additional explanations or examples. • Option to generate a PDF document that summarizes answers / explanations for deeper understanding. • Tracks the topics the student has covered to help tailor responses.

• Face recognition-based attendance: employees' faces are recognized in a specific location to mark presence automatically. • Track number of days present, total working hours per day/week/month. • Leave application & approval workflow. • Pay slip downloads. • Announcements, document request system, calendar of events/holidays, recent activity logs.

• Product inventory management: listing products, tracking stock quantities, adjusting stock after each sale. • Billing / invoice generation when products sold; apply codes, taxes, discounts. • Track product sales over time; generate reports (daily, weekly, monthly). • Predictive insights / suggestions to increase sales: slow-moving products, bundling, special offers. • System to send special offers/promotions to customers.

• A public facing website for the business/agency, showcasing services (e.g. SaaS, web design, graphic design, marketing). • Built with only HTML, CSS, JavaScript, PHP (no external JS/CSS animation libraries). • Animations hand-coded. • Ability for business clients to upload their brand logo + type of business, and in response the system generates a basic website template or poster suited to them. • Contact form that can send messages to their WhatsApp. • Admin dashboard to collect user leads/data, track which pages/services are viewed, track traffic. • Portfolio page that is clean, well laid out, responsive design.

Background: Generative AI is widely used, but prompt injection attacks can trick models into revealing confidential data or bypassing restrictions. Challenge: Build a real-time prompt injection detection & defense system that filters harmful prompts without blocking normal user input. Expected Outcome: A middleware/API wrapper for LLMs that flags or blocks malicious prompts in <200ms. Datasets: • AdvBench – adversarial prompt dataset (https://huggingface.co/datasets/walledai/AdvBench) • JailbreakBench – misuse/jailbreak behaviors (https://jailbreakbench.github.io/) • Safety Prompts Data Catalogue (https://safetyprompts.com/)

Create a tool that scans a given directory or disk image for hidden files, deleted files, or steganography images using basic forensic techniques (e.g., file signature analysis, carving). Objective: Help beginners practice digital forensics skills. Input: Disk image (.img) or folder with files (some intentionally hidden). Output: List of recovered hidden/deleted files and extracted hidden images.

Problem & Challenge: Children in crowded areas risk getting lost. Develop an IoT system that tracks children in real time, detects when they leave safe zones, and notifies guardians instantly. The system should be accurate, low-power, and reliable. Objectives & Scope: • Wearable devices using GPS, RFID, BLE/LoRa for real-time tracking • Geofencing alerts for safe-zone breaches • Mobile/web platform for monitoring and notifications • Optimize low power consumption and validate through simulation and real-world testing

• A public facing website for the business/agency, showcasing services (e.g. SaaS, web design, graphic design, marketing). • Built with only HTML, CSS, JavaScript, PHP (no external JS/CSS animation libraries). • Animations hand-coded. • Ability for business clients to upload their brand logo + type of business, and in response the system generates a basic website template or poster suited to them. • Contact form that can send messages to their WhatsApp. • Admin dashboard to collect user leads/data, track which pages/services are viewed, track traffic. • Portfolio page that is clean, well laid out, responsive design.

Background: Generative AI is widely used, but prompt injection attacks can trick models into revealing confidential data or bypassing restrictions. Challenge: Build a real-time prompt injection detection & defense system that filters harmful prompts without blocking normal user input. Expected Outcome: A middleware/API wrapper for LLMs that flags or blocks malicious prompts in <200ms. Datasets: • AdvBench – adversarial prompt dataset (https://huggingface.co/datasets/walledai/AdvBench) • JailbreakBench – misuse/jailbreak behaviors (https://jailbreakbench.github.io/) • Safety Prompts Data Catalogue (https://safetyprompts.com/)

Problem Statement: Design and implement a secure banking API system that handles real-time fund transfers, account balance checks, and transaction history. The system must implement JWT-based authentication with access tokens (15-minute expiry) and refresh tokens (7-day expiry). Requirements: • Implement user login endpoint that returns both access and refresh tokens • Create protected endpoints for: balance inquiry, fund transfer, transaction history, and beneficiary management • Implement token refresh mechanism without requiring re-login • Handle concurrent transaction requests to prevent race conditions • Implement rate limiting (max 100 requests per minute per user) • Add real-time transaction notifications using WebSockets • Ensure atomic transactions (transfer should debit one account and credit another as a single operation) • Implement token blacklisting for logout functionality • Add 2FA verification for high-value transactions (>$10,000) Bonus Challenges: • Implement token rotation (issue new refresh token on each refresh) • Add device fingerprinting to bind tokens to specific devices • Implement suspicious activity detection (multiple failed auth attempts)

A trustless marketplace for freelancers and clients where smart contracts hold payments and release them based on milestone completion or dispute resolution through decentralized arbitration. • No platform fees (unlike Upwork/Fiverr taking 20%) • Reputation system stored on-chain • Automated payments upon work delivery • Community-based dispute resolution

Theme: Digital Asset Trading Problem Statement: Build a decentralized NFT marketplace where users can mint, buy, sell, and trade digital assets securely using blockchain. Requirements: • Smart contracts for minting and transferring NFTs • User wallet integration (MetaMask or WalletConnect) • Display NFT metadata (image, price, owner, history) • Implement royalties for creators • Transaction history and verification on blockchain Bonus Challenges: • Enable cross-chain NFT trading • Integrate AI for NFT valuation or trend prediction • Add lazy minting for gas optimization

• A public facing website for the business/agency, showcasing services (e.g. SaaS, web design, graphic design, marketing). • Built with only HTML, CSS, JavaScript, PHP (no external JS/CSS animation libraries). • Animations hand-coded. • Ability for business clients to upload their brand logo + type of business, and in response the system generates a basic website template or poster suited to them. • Contact form that can send messages to their WhatsApp. • Admin dashboard to collect user leads/data, track which pages/services are viewed, track traffic. • Portfolio page that is clean, well laid out, responsive design.

Theme: Cyber Awareness & Training Problem Statement: Design an Android app that simulates real-world cybersecurity attack scenarios to train users in identifying phishing, malware, and social engineering threats. Requirements: • Interactive phishing email and message simulations • Score-based evaluation for each simulation • Educational feedback after each attempt • Admin panel for generating new training scenarios • Track user progress and awareness improvement Bonus Challenges: • Integrate AR for physical security awareness scenarios • Implement gamified leaderboard system • Generate phishing templates using AI

• Face recognition-based attendance: employees' faces are recognized in a specific location to mark presence automatically. • Track number of days present, total working hours per day/week/month. • Leave application & approval workflow. • Pay slip downloads. • Announcements, document request system, calendar of events/holidays, recent activity logs.

Theme: Augmented Reality Navigation Problem Statement: Build an Android AR app that helps users navigate both indoor and outdoor environments using real-world overlays and 3D path guidance. Requirements: • Use ARCore for 3D navigation overlays • Integrate GPS for outdoor navigation • Use Wi-Fi or Bluetooth beacons for indoor positioning • Display arrows and labels in AR view • Include search for specific rooms, buildings, or points of interest Bonus Challenges: • Add voice-assisted navigation • Enable multi-floor navigation • Support offline maps and routes caching

Theme: Social Collaboration Platform Problem Statement: Develop an Android application that acts as a digital community hub for clubs or groups, allowing chat, voice, and event collaboration similar to Discord. Requirements: • Create chat channels and voice rooms • Implement real-time messaging with Firebase or WebSocket • User roles and permissions for admins/moderators • Event scheduling and notifications • Support file/image sharing within groups Bonus Challenges: • Integrate AI chatbot for Q&A or moderation • Add voice-to-text transcription for accessibility • Implement message encryption for privacy

Theme: Data Authenticity Verification Problem Statement: Develop a blockchain-based system that verifies the integrity of files by storing their hash values immutably on a distributed ledger. Requirements: • Generate SHA-256 or SHA-3 hash for uploaded files • Record hash on blockchain (Ethereum/Polygon) • Verify file authenticity by comparing hash values • Display transaction logs and verification results • Provide API endpoints for integration Bonus Challenges: • Support batch verification and scheduled checks • Integrate IPFS for decentralized file storage • Implement tamper alert notifications

Theme: Web Security & Threat Detection Problem Statement: Build a machine learning classifier that can identify phishing URLs with high accuracy using publicly available datasets. Requirements: • Feature extraction from URLs (domain age, HTTPS, special characters, etc.) • Train models using multiple algorithms (Random Forest, XGBoost, Neural Networks) • Achieve >95% accuracy on test datasets • Create API endpoint for real-time URL classification • Provide confidence scores and explanation for predictions Bonus Challenges: • Integrate with browser extension for real-time protection • Support multilingual phishing detection • Implement model retraining pipeline with new threat data

Problem: Users often lack awareness of identity theft risks and preventive measures. Additionally, fraud detection is usually reactive, and users have limited exposure to real-time scenarios in banking, e-commerce, and digital identity management. Challenge: Develop an AI-driven platform that detects and prevents identity theft and fraud. The platform should simulate real-time scenarios, including unusual login patterns, fake documents, suspicious transactions, and e-commerce scams. Users should be able to practice AI-guided detection and decision-making while receiving personalized feedback. Deliverables: • Simulated identity theft and fraud cases • Live scenario simulations with AI-guided feedback • Interactive detection and preventive exercises • Decision-making scoring system • Personalized modules and risk assessment quizzes

Theme: Intelligent Network Monitoring Problem Statement: Develop an AI-powered system that monitors real-time network traffic and detects anomalies, potential intrusions, or data exfiltration attempts. Requirements: • Capture live network packets and extract features • Use ML algorithms for anomaly detection • Classify normal vs malicious traffic • Provide dashboards for visualization (traffic flow, protocol stats) • Support alerts on detected anomalies Bonus Challenges: • Integrate with SNMP or NetFlow for enterprise monitoring • Implement explainable AI to justify anomaly decisions • Simulate DDoS or port scan attacks for testing

Theme: Endpoint Security & Response Problem Statement: Design an intelligent ransomware detection system that can identify abnormal file encryption activity and automatically isolate infected endpoints to prevent network-wide spread. Requirements: • Monitor filesystem activity for encryption patterns • Detect abnormal CPU/disk usage and unauthorized file access • Implement automated containment (e.g., network isolation or process kill) • Use ML anomaly detection for real-time ransomware behavior recognition • Provide centralized dashboard for alerts and containment logs Bonus Challenges: • Implement sandbox analysis for unknown executables • Integrate with SIEM systems (e.g., Splunk/ELK) • Simulate ransomware attacks safely for testing detection accuracy

Create a tool that scans a given directory or disk image for hidden files, deleted files, or steganography images using basic forensic techniques (e.g., file signature analysis, carving). Objective: Help beginners practice digital forensics skills. Input: Disk image (.img) or folder with files (some intentionally hidden). Output: List of recovered hidden/deleted files and extracted hidden images.

Anti-counterfeiting system for high-end products (watches, handbags, sneakers) with verifiable authenticity and ownership history. • Each item gets a unique blockchain identity at manufacture • Track entire ownership chain (useful for resale) • Prove authenticity instantly via QR code • Combat fake luxury goods market • Warranty and repair history tracking

Theme: Digital Asset Trading Problem Statement: Build a decentralized NFT marketplace where users can mint, buy, sell, and trade digital assets securely using blockchain. Requirements: • Smart contracts for minting and transferring NFTs • User wallet integration (MetaMask or WalletConnect) • Display NFT metadata (image, price, owner, history) • Implement royalties for creators • Transaction history and verification on blockchain Bonus Challenges: • Enable cross-chain NFT trading • Integrate AI for NFT valuation or trend prediction • Add lazy minting for gas optimization

Theme: Intrusion Deception Systems Problem Statement: Develop a honeypot system that detects and logs unauthorized access attempts, analyzes attacker behavior, and sends alerts to administrators. Requirements: • Deploy multiple honeypots (SSH, HTTP, FTP, etc.) • Log attacker commands, IP, and timestamps • Analyze attack patterns and generate behavior profiles • Alert administrators in real time via dashboard or email • Implement visualization for attack trends Bonus Challenges: • Integrate ML for attack classification (bot, script, human) • Generate fake data environments to mislead attackers • Correlate honeypot data with firewall logs

Background: Deepfakes are used in fraudulent banking calls, KYC bypass, and misinformation campaigns. Current detection is mostly offline. Challenge: Build a real-time multimodal deepfake detector (voice + video + behavior analysis). Expected Outcome: Verification layer for live video calls or financial KYC workflows. Datasets: • Face Forensics++ – https://github.com/ondyari/FaceForensics • ASVspoof 2019 – https://www.asvspoof.org/ • Deep Fake Detection Challenge (DFDC) – https://ai.meta.com/datasets/dfdc/

Theme: Social Collaboration Platform Problem Statement: Develop an Android application that acts as a digital community hub for clubs or groups, allowing chat, voice, and event collaboration similar to Discord. Requirements: • Create chat channels and voice rooms • Implement real-time messaging with Firebase or WebSocket • User roles and permissions for admins/moderators • Event scheduling and notifications • Support file/image sharing within groups Bonus Challenges: • Integrate AI chatbot for Q&A or moderation • Add voice-to-text transcription for accessibility • Implement message encryption for privacy