Darshan B.

Key Skills

Software

No software listed

Top Subject Matter

E-commerce

Finance and Business Analystics

HealthCare

Top Data Types

Text

Top Task Types

Classification

Freelancer Overview

I have experience working on AI and machine learning projects involving data preprocessing, labeling, sentiment classification, and model evaluation. One of my major projects was a real-time YouTube live sentiment analysis system that analyzed live stream comments and classified audience sentiment as positive, negative, or neutral. In this project, I worked extensively with NLP techniques such as text cleaning, tokenization, TF-IDF vectorization, and dataset preparation for training machine learning models. I trained and evaluated multiple algorithms including Support Vector Machine (SVM), Random Forest, and Naive Bayes to improve prediction accuracy and real-time performance. The project was integrated with a live dashboard using Python and Streamlit for continuous sentiment monitoring and visualization. In addition, I have published a research paper titled Techniques of Ensemble Learning for Human Emotional Classification and Detection in the International Journal of Scientific Research in Engineering and Management, which focused on ensemble learning techniques for emotion detection and classification. This research strengthened my understanding of AI training workflows, dataset preparation, annotation quality, and model optimization. My technical background in Python, NLP, machine learning, and full-stack development enables me to efficiently work with AI training data, validate outputs, and improve data quality for better model performance.

Labeling Experience

Youtube Live Sentiment Analysis

TextClassification

Developed a real-time YouTube live sentiment analysis system that collects and analyzes live stream comments using Natural Language Processing (NLP) and Machine Learning techniques. The project processes live comments in real time, performs text preprocessing such as tokenization, stop-word removal, stemming, and TF-IDF vectorization, and classifies sentiments into positive, negative, and neutral categories. Multiple machine learning models including Support Vector Machine (SVM), Random Forest, and Naive Bayes were trained and evaluated to achieve higher prediction accuracy and efficient sentiment detection. The system was built using Python and Streamlit to create an interactive real-time dashboard displaying audience sentiment trends, live comment analysis, and graphical insights. The project focused on improving model performance through proper dataset preparation, feature extraction, and ensemble learning approaches. This project enhanced my expertise in AI model training, data preprocessing, sentiment classification, NLP workflows, and real-time analytics systems.

2024 - 2025

Education

Work History