Track
Dive into the world of artificial intelligence by building models that can learn and make decisions. Explore data processing, neural networks, and real-world AI applications.
Start here
Develop a binary classification model to predict the onset of diabetes based on diagnostic medical measurements. This is a foundational project for working with real-world, tabular health data and understanding the impact of different features on a model's prediction.
Sign in to track progress
Basic
4 projectsDevelop a binary classification model to predict the onset of diabetes based on diagnostic medical measurements. This is a foundational project for working with real-world, tabular health data and understanding the impact of different features on a model's prediction.
Build a system that automatically classifies music tracks into different genres (e.g., Blues, Classical, Hip-Hop, Rock) based on their audio features. This project is an excellent introduction to working with audio data and signal processing.
Build a model to classify text reviews (e.g., from movies or products) as positive or negative. This is a foundational project in Natural Language Processing (NLP) that serves as an excellent introduction to text data.
Develop a regression model to predict housing prices based on various features like area, number of bedrooms, location, and age. This project introduces core machine learning regression concepts and data analysis.
Intermediate
5 projectsDevelop a deep learning model that can detect and classify traffic signs from images or a video stream. This project is a practical application of computer vision and is a fundamental component for autonomous driving systems.
Create a recommendation engine that suggests movies to users based on their past ratings. You can implement both collaborative filtering (user-user or item-item) and content-based filtering, providing a comprehensive understanding of recommender systems.
Build a time-series model to forecast future energy consumption for a household or a region. This project is highly relevant for smart grid management, resource allocation, and sustainability efforts, and provides deep experience in handling temporal data.
Create a robust system to detect fraudulent credit card transactions from a real-world dataset. This project is highly practical and focuses on the critical challenge of working with highly imbalanced datasets.
Build a model to predict which customers are likely to cancel their subscription to a service (e.g., a telecom or SaaS company). This model provides direct business value by enabling proactive customer retention strategies.
Advanced
4 projectsCreate a model that combines the content of one image with the artistic style of another. This project provides a fascinating look into the inner workings of Convolutional Neural Networks and how they encode content and style information in different layers.
Develop a deep learning model that generates a short, coherent, and fluent summary of a long text document (like a news article). Unlike extractive methods that pick sentences, this model will generate new sentences, demonstrating a deeper understanding of the text.
Develop a Generative Adversarial Network (GAN) to generate novel, realistic images in a specific domain (e.g., human faces, anime characters, or modern art). This project delves into the creative and complex world of generative models.
Build a deep learning model that transcribes spoken audio into text. This is a challenging sequence-to-sequence problem with significant real-world applications like voice assistants, automated subtitling, and meeting transcription.
Expert
4 projectsTake a large, high-performance deep learning model and optimize it to run efficiently on a resource-constrained device (like a smartphone or Raspberry Pi). This MLOps-focused project bridges the gap between model development and real-world production deployment.
Train a model to learn meaningful visual features from a massive, unlabeled image dataset. The goal is to create a powerful feature extractor that can then be fine-tuned for various downstream tasks (like classification or detection) using very little labeled data. This project explores the cutting-edge of reducing reliance on expensive human annotation.
Design and train a Reinforcement Learning (RL) agent that learns an optimal trading strategy for a portfolio of financial assets. The agent will decide when to buy, sell, or hold assets to maximize a risk-adjusted return metric like the Sharpe ratio.
Implement a federated learning system to train a global model on data distributed across multiple clients (simulating, for example, mobile devices) without the raw data ever leaving those clients. This project tackles the critical real-world challenge of building AI systems that are private-by-design.
