May 2025 - June 2025 | 40 days

• Research on predicting mortality using mRS in Intracerebral Hemorrhage patients
• Predicted Intracerebral Hemorrhage (ICH) patient mortality using mRS scores (0-6 scale)
• Preprocessed hospital data (under NDA) using TVAE, Bootstrap, and SMOTE, reducing error by 30-35%
• Benchmarked 10+ models including Random Forest, LightGBM, CatBoost, Naive Bayes, and others

June 2024 | 3 Weeks | Organized by Corporate Gurukul in collaboration with AWS

Curriculum Highlights:

• Training in Machine Learning and Deep Learning by NUS professors
• Hands-on AWS sessions covering cloud tools, model deployment, and scaling

Achievements: Ranked in top 10% in academic assessments

August 2023 - January 2024 | 6 Months

• Designed logos, event posters, social media creatives, and brand identity materials
• Developed branding concepts using mood boards, color palettes, typography, and layout principles
• Find my Conceptual Branding Graphic Design Portfolio that got me this role by clicking on the image.

• Capstone Project of GAIP internship (National University of Singapore)
• Used Indian Pines & Salinas datasets
• Applied 3D CNN and ResNet models
• Achieved 99% classification accuracy
• Focused on spectral band preprocessing, dimensionality reduction, and model evaluation
• Highest scoring team project in the batch
• Unprecedented accuracy in the past 8 years of GAIP

• Static clone of YouTube homepage using HTML and CSS
• Responsive design using Flexbox and Grid
• Interactive elements with hover effects and tooltips
• Modular CSS structure for maintainable code

• Full-featured e-commerce interface with product browsing and cart management
• Dynamic cart updates with quantity adjustment and item removal
• Checkout flow with delivery options and real-time price calculations
• Persistent user data using Local Storage API

• Full-stack web application to review, rate, and discover books
• Features user authentication, single review per book, and half-star ratings
• Displays dynamic average ratings and top reviewed books updated in real-time
• Content-based recommendation system leveraging author, genre, keywords, and themes
• Interactive AI mascot enhances user experience with guided messaging
• Currently used by a small user base of friends and connections

• Content-based filtering system using cosine similarity
• User inputs favorite movie, system returns similar recommendations
• Refined output from top 50 to top 30 based on user feedback
• Designed and implemented UI using Streamlit

• Developed CNN architecture for recognizing handwritten numerals
• Datasets included Urdu (Kaggle dataset) and other regional Indian language datasets
• Implemented preprocessing pipeline with Gaussian blur and Otsu's thresholding
• Deep CNN architecture with multiple convolutional layers and strategic pooling

• 10-second 3D animation clip created in Blender
• Designed donuts from scratch starting with torus mesh
• Added detailed glazing and sprinkles
• Used Poliigon assets for background elements
• 180 frames rendered at 24fps (~3 days render time)

• Hybrid LSTM-BERT model achieving 97% accuracy
• Outperformed standalone models in detecting sophisticated misinformation
• Presented onsite at CODE AI 2025 (Manipal Dubai)
• Accepted for publication in Springer proceedings

• Built the image classifier using ResNet18 with 97.5% accuracy on CIFAKE dataset
• Comparative study of VGG16 vs ResNet architectures
• Full pipeline from data sourcing to model evaluation documented
• Presented onsite at ETCC 2025 (PES Bengaluru)
• Accepted for publication in IEEE Xplore

• Systematic analysis of 90+ papers across 20 years
• Novel classification into Primitive (2000-2010), Transitional (2011-2016), and Modern (2017+) eras
• Compared AI effectiveness in mitigation, preparedness, response, and recovery phases
• Submitted to Natural Resources Forum journal

• Executed 3 major events with 60-120+ participants each
• Organized Industry Connect with Google speaker for career guidance
• Conducted STEM outreach at government school for International Girls and Women in Science Day
• Hosted panel discussion where female interns shared their journey
• Published 2 LinkedIn articles and tripled social media engagement