Full Stack and UI/UX Engineer. I build scalable user centric platforms

productHunt

gumroad

leetcode

github

Aryan Lodha  • Peerlist

EcoVision is an AI-powered platform tackling smart city environmental management by integrating real-time computer vision, crowd analytics, and sustainability metrics into a single responsive dashboard. It enables: 🗑️ Automated Trash Segregation & Waste Detection 🧍 Real-Time Crowd Density Mapping & Flow Tracking 🌱 Carbon Footprint Estimation based on live activity 📊 Dynamic Visual Dashboards for decision-making Why it stands out: Built in under 24 hours during BrisHack 2025, EcoVision won the People’s Choice Award for its innovation, functionality, and real-world applicability.

EcoVision - AI-Driven Smart Waste & Crowd Man

Built an end-to-end pipeline to classify cognitive load using instruction-free eye-tracking data. Engineered 150+ gaze features (fixations, saccades, blinks, entropy, pupil metrics). Benchmarked ML models, achieving 93% accuracy with Random Forest. Developed a Hybrid TCN + ViT deep learning model, achieving 98.8% accuracy & 0.99 F1-score. Applied SHAP and Grad-CAM++ to interpret feature importance and temporal attention. Validated generalisation using subject-wise splits across 52 participants (196 + 80 trials).

Evaluating the Importance of Time Series Feat

Objective: Seamless, user-friendly product sharing with a focus on convenience, sustainability, and trust. Key Features: Automated subscription models Advanced search filters Robust return policies Verified product authenticity User-friendly interface Design Process: Empathy mapping User personas Iterative prototyping User testing

GoCart - Convenience at your fingertips

Knee arthritis is a significant concern affecting patient health, and early prediction is crucial for effective management. This project explores the use of deep learning models, including VGG16, ResNet50, DenseNet, and EfficientNetB5, for knee arthritis severity assessment. The most promising model, EfficientNetB5, was selected and further improved through image augmentation techniques, resulting in a remarkable accuracy improvement of 97%.

Image Augmentation Upshot In Enhanced Knee Ar

Processed 1.97M+ morphometric records and integrated 377+ datasets with CENOGRID climate proxies to build a unified analytical pipeline. Analyzed community structure using KDE, PCA, skewness/kurtosis, and entropy, revealing major morphological regime shifts over time. Identified key evolutionary transitions at ~3.2 Ma and ~1.6 Ma using PELT change-point detection, aligned with glacial cycle intensification. Estimated species richness trends using GMM + BIC, highlighting diversification peaks and climate-driven declines. Modeled climate–morphology links via mixed-effects models and lagged correlations, showing strong environmental influence on shape traits.

Patterns of Morphological Diversification in 

• Convenient daily journaling and integrated to-do lists. • Secure storage of data in MongoDB. • Fast and reliable performance with NodeJS/ExpressJS backend. • User-friendly interface built using React.

PLANIT NOW - Complete Productivity Platform

Property Portal is a team project made by me and my friends during a Hackathon we participated in. We made this portal in 24 hours including the time to decide the problem statement and how the portal will work. The project is an online platform designed to revolutionize the way property transactions are conducted. By removing the need for middlemen such as brokers and agents, the platform provides a more direct, cost-effective, and secure means of buying and selling properties. Users can create an account on the platform after undergoing a thorough verification process. To sell a property, the user is required to fill out a form providing all the necessary details. The government officials then verify the information and approve the sale request. All money transfers are processed through the platform, ensuring transparency and security. By leveraging the power of technology and eliminating intermediaries, the platform offers a fast, reliable, and efficient way to conduct property transactions.

Property Portal

This study focuses on the application of the Support Vector Regression (SVR) model, an artificial intelligence-based model, for flood prediction in the Hoppers Crossing catchment in Australia. SVR models have gained popularity for their ability to simulate the instantaneous and nonlinear nature of riverine floods. The model utilizes lagged average rainfall and runoff as input variables, determined by the time of response concept.

Real Time Flood Prediction Through Support Ve

The key role in the redevelopment of the RAJ-SIMS portal, focusing on enhancing supply chain management for the Government of Rajasthan. The main task involved utilizing .NET Core for backend development, Oracle 19c for database management, and React.JS for the front end. Key enhancements included implementing Swagger for API documentation and Jenkins for continuous integration, significantly streamlining deployment processes. These improvements resulted in greater transparency, operational efficiency, and an enhanced user experience across various devices.

Revamping Rajasthan Supply & Inventory Manage

This review paper is about NFT Security how NFTs have emerged as a popular method of financing in digital art and what security measures should be kept while using it to collect and finance digital art.