Watch Brand Classification Model Enhancement

This project aims to enhance the predictive accuracy of an existing watch classification model deployed in AWS SageMaker. The current model, while functional, requires optimization to improve its ability to accurately identify watch brands from images. By utilizing an open-source dataset of watch images, learners will apply advanced machine learning techniques to fine-tune the model. The goal is to develop a refined model prototype that demonstrates a significant and quantifiable improvement in brand identification accuracy. The project will be conducted entirely within the AWS SageMaker environment, providing learners with hands-on experience in a real-world cloud-based machine learning platform. This project offers an excellent opportunity for learners to apply their classroom knowledge of machine learning and cloud computing to a practical problem.

Analytics Project | Gluu

Main Goal for This Project This project will apply data analytics tools and methodologies, centered around PostHog, to build, assess, and optimize the Gluu Admin Dashboard. The objective is to deliver actionable insights based on user data, content metrics, and onboarding funnels to enhance administrative decision-making and improve overall platform performance. Project Overview & Objective Statement Gluu is shifting its focus to develop a robust Admin Dashboard, a critical internal tool for managing the platform and its users. The key objective is to create a centralized interface that provides deep insights into user activity and content engagement. Objective: Analyze user behaviour, content trends, and system-wide statistics to: ● Identify user engagement patterns and pain points within the platform. ● Quantify the effectiveness of the user onboarding process. ● Recommend data-backed improvements for user management and content strategy. ● Create dashboards for monitoring key platform metrics.

Enhancing Peer-to-Peer Rental Transactions for Gluu.Repair

Gluu.Repair is transitioning from Version 1 to Version 2, aiming to enhance its Peer-to-Peer Rental Feature. The current system allows users to share specific collections with clients, but lacks a streamlined process for clients to select items, choose rental durations, and view the total value of their rental requests. The project focuses on designing a comprehensive transaction flow that addresses these gaps. Key tasks include developing a user-friendly interface for clients to select items and rental periods, integrating a system to display the total rental value, and creating a seamless approval and payment capture process. The goal is to ensure a smooth and efficient rental experience for both users and clients, aligning with Gluu.Repair's organizational change objectives.

Image Recognition Model Enhancement for Condition Reporting

The project focuses on enhancing Gluu's multi-modal image recognition model to improve the accuracy and efficiency of condition reporting for client products. The goal is to develop a custom Convolutional Neural Network (CNN) and integrate You Only Look Once (YOLO) object detection techniques to better identify and classify product conditions. This project will allow learners to apply their knowledge of machine learning, computer vision, and neural networks to a real-world application. By focusing on CNN and YOLO, the project aims to create a robust system capable of handling diverse product images and providing precise condition assessments. The project will involve tasks such as data preprocessing, model training, and performance evaluation, all of which are crucial for developing a reliable image recognition system.