Tracks & Themes

Smart cities rely heavily on interconnected IoT devices to manage critical infrastructure, including traffic control, surveillance networks, energy grids, and public safety systems. However, because these devices often operate with weak security configurations, they serve as easy targets for cyberattacks such as Distributed Denial of Service (DDoS), botnets, and unauthorized remote access. A successful attack can paralyze public services and compromise citizen safety. The challenge is to design and develop an intelligent, real-time cyber threat monitoring and automated response framework capable of securing a smart city’s IoT ecosystem.

Develop an AI-powered software solution for automated structural health monitoring. The system should utilize deep learning to accurately detect, localize, and assess the severity of cracks on concrete structures (such as bridges, dams, and buildings) under diverse environmental conditions. Participants are challenged to build a robust pipeline capable of processing image or live video inputs in near real-time, optimized for deployment on resource-constrained edge devices or UAVs. The final solution must feature an intuitive dashboard that visualizes the detected anomalies and provides actionable structural assessments.

Develop a functional software prototype of an intelligent ATS that automatically flags resume anomalies, plagiarism, and AI-generated content in real time. The system should feature a semantic verification engine that cross-checks a candidate's claimed technical skills against their actual project descriptions and portfolio data to detect fraud or inflation. Additionally, the final solution must include an intuitive HR dashboard that displays a detailed fraud risk score for each applicant, providing clear explanations for the assessment results to ensure a fair, transparent, and highly secure technical hiring process.

Develop a functional software prototype that automatically processes digital text inputs of student answers and assigns accurate, objective grades based on semantic and contextual analysis. The system should feature an NLP evaluation engine capable of identifying key concepts, analyzing syntax, and detecting plagiarism or copied content across submissions. Additionally, the final solution must include an intuitive evaluator dashboard that displays the assigned scores, highlights specific gaps or errors in the student's answer, and provides a clear breakdown of the grading rationale for educators to review.

To deliver a functional software prototype consisting of a mobile-responsive application or dashboard that seamlessly manages the rental lifecycle from both the farmer and CHC operator perspectives. The platform should feature an active hyperlocal marketplace with geospatial routing, a dynamic pricing engine that adjusts rates based on live demand variables, and a preliminary agri-fintech credit scoring model for deferred payments. Additionally, the final solution must include an operational management dashboard for CHC operators that simulates IoT telematics data streams (tracking location, usage hours, and fuel consumption) alongside automated billing and predictive maintenance schedules.

To deliver a functional software prototype featuring an AI-driven remote sensing pipeline that processes satellite imagery to calculate vegetation health indices and predict weather-induced yield impacts. The system should incorporate a cryptography layer that successfully generates Zero-Knowledge Proofs, allowing a farmer to securely prove their eligibility for a payout to an insurer while keeping private parameters completely hidden. Additionally, the final solution must include a tamper-proof ledger or blockchain-based log to maintain an unalterable claim trail alongside an intuitive predictive risk dashboard that alerts farmers to potential environmental hazards before they occur.

To deliver a functional software prototype centered around a deep learning computer vision model that accurately classifies various plant diseases from uploaded or live-captured leaf images. The system must process these images in real time and automatically fetch precise, localized treatment recommendations, pesticide dosages, or natural remedies. Additionally, the final solution must feature a simple, multilingual mobile or web-based user interface designed specifically for easy navigation by farmers, alongside an offline-first caching mechanism to ensure usability in low-connectivity rural areas.

A functional web application prototype that processes an uploaded video file or live audio stream. The system must feature an NLP semantic translator that converts standard syntax into Sign Language grammar strings, driving a fluid, animated digital human avatar on-screen with synchronized facial expressions and hand movements to ensure deep comprehension for the viewer.

To develop a functional software prototype, such as a browser extension, a web application plugin, or an API gateway. The system must feature an NLP classification pipeline that ingests digital articles or text links to output a multi-dimensional "Credibility Score Card," highlighting specific unverified claims and emotional clickbait word ratios. Additionally, the solution must incorporate a computer vision module that analyzes visual media for deepfake inconsistencies, paired with an interactive network graph visualization that traces exactly how the misinformation propagates across digital networks.

Deliver a functional software prototype centered around a deep learning pipeline capable of processing and analyzing medical image inputs in real time. The core engine must automatically classify the scan, detect anomalous regions, and generate bounding boxes or heatmaps pinpointing the exact location of potential pathology. Additionally, the final solution must feature a secure clinician-facing web portal that displays the patient's scan alongside the AI's diagnostic confidence score, an automated severity assessment, and a structured, draft clinical report to optimize the radiologist's workflow.