Chinmay Pagey

Education

MTech Integrated Computer Science with specialization in Business Analytics

CGPA - 9.03

Chennai | October 2020 - June 2025

Data Analyst (Internship)

Mumbai | Jan 2025 - Present

Leading the adoption of NexusAI, Crompton's in-house Al platform, now used by 3000+ employees, including leadership.
Improving business workflows by 20% through machine learning model integration, use case definition, and system optimization.
Developing AI tools using Retrieval-Augmented Generation (RAG) to enhance document retrieval, reducing search times by 50% and increasing productivity.
Contributing to the development of HR Policy Assist and document query systems, improving self-service rates by 35% and automating internal processes.
Collaborating on factory automation using YOLOv9 to detect missing components in packaging, contributing to quality control.
Working with large datasets to extract insights that support data-driven decision-making across teams.

Data Analyst (Internship)

Bangalore | May 2023 - August 2023

Improved clients' marketing strategies, resulting in a 35% increase in audience reach through data-driven insights.
Engineered an automated data extraction process using Apollo CRM, enabling seamless delivery of 5000+ verified contact details to clients monthly, facilitating targeted outreach and boosting their lead conversion rates by 40%.

SEO team (Internship)

Mumbai | May 2022 - July 2022

Supported the SEO team by researching and recommending keywords for the newly launched website 'Good Health by Yourself', leading to a 400% increase in web traffic within a short period.
Gained hands-on experience with Semrush and Ahrefs to identify high-impact, relevant keywords across 9+ countries, enhancing the website's global search visibility and content relevance.

Deep Learning

Achieved 98.72% accuracy by developing OncoCutisAI, a hybrid diagnostic model combining 17 metadata features and 10,000+ augmented dermoscopic images using ensemble learning with Swin-Tiny, MobileViT, and XGBoost.
Preprocessed data for 1,373 patients, using MICE to impute 15% missing values and applying 5 augmentations per image to enhance deep learning model generalization.
Benchmarked multiple architectures (ResNet, Efficient Net, MobileNet) and ensemble techniques (majority voting, weighted averaging) to optimize performance and reliability.

Natural Language Processing

Developed an automated essay grading system using Bi-LSTM architecture, achieving 96% accuracy in grading cohesion, syntax, and vocabulary for essays by English learners (Grades 8-12).
Incorporated BERT for plagiarism detection by computing a similarity index, comparing essays to detect potential copying.

Computer Vision

Developed Nivaaran for SIH Hackathon, a smartphone app that integrates Instagram and Twitter APIs to collect real-time data on water concerns.
Applied MobileNet for flood image detection with an accuracy of 93%, enhancing its capabilities to effectively address water-related issues in communities.
Integrated the model into the app to detect flooded areas, classify water-related issues from social media images, and assess flood risk for timely disaster response.