Rohan Badiger
Development
Maharashtra, India
Skills
Machine Learning (ML)
About
Rohan Badiger's skills align with IT R&D Professionals (Information and Communication Technology). Rohan also has skills associated with System Developers and Analysts (Information and Communication Technology). Rohan Badiger appears to be an entry-level candidate, with 11 months of experience.
View more
Work Experience
Image Processing Intern
Centre for Development of Advanced Computing
June 2023 - Present
- Optimized algorithm versions and utilized deep learning techniques to develop and evaluate a high-performing model. • Gathered a diverse datasets of 20,000 images across 40+ classes for model training. • Achieved 85% accuracy through optimized algorithms and advanced deep learning techniques, showcasing the model's precision and effectiveness. • Leveraged data augmentation techniques to address insufficient data, enhancing the model's robustness and performance in scenarios with limited available data. • Enabled real-time model detection for images, videos, and live camera feeds, enhancing practical applicability.
Mentor and Tester
DeepLearning.AI
May 2023 - Present
- Remote • As an Alpha tester for the "Generative AI with LLMs" course, I play a pivotal role in refining the curriculum by identifying and addressing issues. • Simultaneously, served as a mentor to guide and support learners on concepts and facilitating a successful learning journey. • Established and maintained an official account on the DeepLearning.AI community website to actively engage with learners, address their queries, and foster a sense of community among participants. • This allowed me to contribute to the course's continuous improvement while providing guidance to learners and addressing their doubts. Projects Visual Object Detection • The project recognizes and locates 600+ custom objects in a scenario with accuracy around 80%. • The project leverages YOLOv8 (You Only Look Once) algorithm for real-time object detection across different labels. • Utilized Streamlit for a user-friendly interface, enhancing the overall system interaction. Clothing item Classifier • Designed and implemented a clothing item detection algorithm using a Convolutional Neural Network (CNN), achieving an impressive 90% accuracy across 10 different classes of clothing items. • Collected and processed data from diverse sources, predominantly public, open-source datasets, and platforms such as Kaggle. • Conducted thorough preprocessing, including data cleaning, outlier removal, image resizing, normalization, and occasionally creating a validation set for fine-tuning hyperparameters. Handwritten Digit Detection • Emphasized the practical applications of recognizing handwritten digits in various domains. • Utilized the MNIST dataset as a benchmark for developing and evaluating recognition algorithms. • Implemented front end system allowing users to draw digits for instant recognition feedback, ensuring interactive experience.
Education
