Experience

Jan 2023 - March 2023

• Designed and developed an advanced multi-class sentiment analysis model capable of predicting over 28 distinct emotions, extending beyond traditional polarity-based sentiment (positive, negative, neutral).
• Built a large-scale, balanced dataset consisting of 111,866 labeled training samples, 11,198 validation samples, and 5,598 test samples, drawn from reviews (Amazon, Flipkart, ProductHunt) and social media feeds (Twitter), with additional augmentation via synthetic data generation and AI-assisted labeling.
• Conducted model experimentation beginning with baseline methods such as FastText and exploratory use of LLMs for zero/few-shot classification to establish benchmarks.
• Advanced to transformer-based architectures, fine-tuning both BERT-base and RoBERTa-base models using Low-Rank Adaptation (LoRA) and custom classification layers to optimize contextual emotion detection.
• Explored alternative modeling strategies such as masked language modeling for emotion inference, where blanks are left within text, where model supposed to predicted the right emotion tokens. This approach showed promise on training data but underperformed on validation and test sets compared to direct classification.
• Achieved classification accuracy exceeding 85% on held-out test sets, demonstrating strong generalization and outperforming traditional models by a significant margin.

June 2024 - July 2024

• a data science internship focused on performing descriptive analysis and deriving actionable insights from multiple structured datasets.
• Worked with three distinct datasets and developed interactive dashboards using the Plotly library in Python, providing clear visualization of key findings.
• Conducted a bird strike analysis, uncovering patterns related to strike occurrences during takeoff, landing, and in-flight conditions, and examined correlations with flight altitude and atmospheric factors.
• Performed a heart disease dataset analysis, identifying major health-related features correlated with heart disease and quantifying relationships between patient attributes and disease occurrence.
• Delivered an employee attrition analysis, analyzing workforce data to detect correlations between job-related factors (salary, work-life balance, tenure) and attrition rates.
• Designed and deployed all three dashboards to Heroku, enabling interactive access to insights and facilitating decision-making through visualization-driven storytelling.

August 2023 - May 2025

• Pursued a Master's degree in Computer Science at the University of Madras, Chennai, actively engaging in diverse academic projects spanning software development, machine learning, embedded systems, and data science.
• Developed a Twitter clone application with a React.js frontend and a Java Servlet backend, leveraging Maven for project management and build automation.
• Created a Python-based word guessing game given a riddle, scoring mechanisms, and user interactivity to demonstrate applied programming concepts.
• Designed and implemented an embedded systems project using Arduino, integrating a temperature sensor and pulse sensor to measure physiological parameters and display real-time outputs on an LCD module.
• Built an image compression model using Variational Autoencoders (VAEs), experimenting with lossy compression techniques and validating theoretical concepts of neural network–based compression.
• Conducted time series analysis & forecasting , applying multiple predictive models to different datasets to analyze trends and generate future projections.
• Implemented a image classification project for brain tumor detection, where ensemble of models are used includin a custom built CNN network and fine tuning state of the art CNNs, achieving an accuracy of 95%

Mar 2023 - May 2023

• a full-stack development internship, contributed to three major projects: a tourism portfolio platform, an ed-tech application, and an e-commerce platform.
• Utilized a diverse technology stack including Python, JavaScript, TypeScript, React.js, Next.js, React Query, Tailwind CSS, Express.js, Nest.js, and FastAPI, with MongoDB as the primary database.
• In the tourism portfolio project, resolved critical hydration errors in the Next.js frontend and conducted API testing to ensure functionality and stability.
• In the ed - tech project, implemented frontend wireframes, developed reusable React components, integrated React Query for API communication, and built backend logic in Nest.js, including controllers, database connections, and API routes.
• In the e- commerce project, worked extensively on the backend with FastAPI, implementing optimized API routes, ensuring efficient database interactions, and integrating caching mechanisms to minimize response times.
• Applied best practices in logging, performance optimization, and scalability, ensuring APIs were production - ready, efficient, and maintainable.

August 2020 - May 2023

• Completed a Bachelor’s degree in Computer Science at MIT World Peace University, Pune, Maharashtra, engaging in multiple academic and self-learning projects in full-stack development.
• Designed and developed UConnect (University Connect), a content management system (CMS) serving as a proof-of-concept platform for communication between students, faculty, and university departments. Built with TypeScript across both frontend and backend. Implemented the frontend in React.js, providing interactive UI components. Developed the backend using AWS Lambda functions and integrated with DynamoDB for scalable, serverless data storage.Enabled users to log and manage complaints with specific departments, ensuring systematic resolution and tracking.
• Supplemented learning with several mini-projects in frontend and backend development, including building portfolio sites, web pages, and API implementations, reinforcing practical software engineering skills.