Divyey Arora

Divyey Arora

About

Detail

AI/ML Intern at Kaay Labs - Software Development
Bengaluru, Karnataka, India

Contact Divyey regarding: 
work
Full-time jobs

Timeline


work
Job
school
Education
folder
Project

Résumé


Jobs verified_user 0% verified
  • K
    AI/ML Intern
    Kaay Labs Software Development
    Mar 2025 - Current (1 year 5 months)
  • Mindera
    Software Developer and AI/ML Intern
    Mindera
    Nov 2024 - Mar 2025 (5 months)
  • C
    Data Scientist
    Cranes Varsity Pvt Ltd
    Oct 2023 - Dec 2023 (3 months)
Education verified_user 0% verified
  • MVJ College of Engineering Bangalore India
    Bachelor of Engineering - BE, Artificial Intelligence and Machine Learning (AI&ML)
    MVJ College of Engineering Bangalore India
    Jan 2021 - Jan 2025 (4 years 1 month)
    Activities and Societies: Hosted an AI workshop for 2nd & 3rd year students (collaboration with IIT Ropar x masai x Tensor AI x MVJCE) (Head of Technical Team, Tensor AI). Head of Organizing the MVJ Football Tournament (MVJFC). GDSC - Google Development Software Club - participated in various ML workshops.
Projects (professional or personal) verified_user 0% verified
  • A
    AI Resume Sceening System (RAG-LLM-AI)
    Feb 2025
  • A
    A/B Testing for Website Higher Engagement/Conversion Optimization
    Dec 2024 - Feb 2025 (3 months)
    Objective: This project aimed to determine whether implementing a dark mode version of a website would lead to higher user conversion rates compared to the existing light mode. Approach: Conducted an A/B test to measure the impact of dark mode on user conversion rates, analyzing 286K+ sessions. 1. Data Cleaning & Preparation: Removed 3,894 duplicate user sessions to ensure data integrity. 2. Sampling Strategy: Used statistical power analysis to determine an optimal sample size of 4,720 users per group (9,440 total). 3. A/B Testing: Conducted a proportion z-test to compare conversion rates between control (light mode) and treatment (dark mode). 4. Statistical Significance: Set α = 5% and test power = 80% to detect a meaningful 2% effect size
  • A
    AI-Powered World Cup Match Outcome Prediction
    Jun 2024 - Nov 2024 (6 months)
    Techniques Applied: Python, Scikit-learn, MLPClassifier, LIME, SHAP. Problem Statement: Predicted outcomes of World Cup Football matches using machine learning models based on historical data to predict match outcomes. Solutions: 1. Collected and preprocessed historical match data, including team stats, player performance, and match outcomes (Data Scraping from websites and Data Cleaning). 2. Engineered features such as possession percentage, shots on target, and defensive metrics. 3. Built a MLPClassifier-based model to predict match results (win, draw, loss).Trained multiple machine learning models and optimized hyperparameters for better accuracy. 4. Utilized LIME and SHAP for explainability, identifying key factors influencing predictio
  • C
    Customer Segmentation of Bank Transactions Data using Machine Learning
    Customer Segmentation of Bank Transactions Data using ML - K-means, PCA & Autoencoders In the competitive landscape of financial services, understanding customer behaviour and preferences is crucial for designing effective marketing strategies, enhancing customer satisfaction, and driving profitability. Financial institutions face the challenge of analysing and segmenting large volumes of diverse customer data, which can hinder their ability to respond effectively to market demands. One primary issue is the presence of heterogeneous customers. Financial institutions encounter wide variations in demographics, behaviours, and financial needs among their clientele. This diversity makes it difficult to develop one size-fits-all solutions, as di