Hi, I am Pratik Thorat
I'm the Data Analyst you need.

AtliQ Hardwares, a fast-growing consumer electronics company, lags in leveraging data for competition due to reliance on Excel for most reports. My Objective: Implement an advanced Power BI analytics solution for insightful decision-making and a competitive edge.

In this project, I took the lead in enhancing English to Hindi translation, resulting in a remarkable 20% improvement in translation quality, as measured by a BLEU score of 0.85. I also implemented efficient deployment processes by leveraging GitHub Actions, which significantly reduced deployment time by 30%. This project involved the use of various technologies, including Python, the Hugging Face Transformers Library, and AWS services such as EC2 and ECR, all seamlessly integrated with GitHub Actions.

In one of my most exciting projects, I successfully developed a real-time sign language gesture recognition system operating at an impressive 30 frames per second (FPS) using YOLOv5. This cutting-edge technology allowed me to achieve an outstanding 90% accuracy in detecting sign language phrases, making it a highly effective communication tool. To make it accessible to a wider audience, I deployed the system as a user-friendly web application, ensuring that the benefits of this innovative solution could be enjoyed by many.

I constructed a predictive model that anticipates air pressure sensor failures in heavy-duty vehicles, resulting in substantial annual cost savings of $500,000. Employing a combination of Logistic Regression, Random Forest, Decision Tree, and XgBoost algorithms, the model achieved an impressive F1 score of 0.87. To enhance efficiency, I streamlined deployment processes using GitHub Actions, reducing deployment time by 40%. In this project, I leveraged various technologies including Python, the Transformers Library, scikit-learn, AWS (EC2, ECR), GitHub Actions, and Docker to deliver a comprehensive and effective solution.