From the course: Advanced Data Processing: Batch, Real-Time, and Cloud Architectures for AI
Join today to access over 24,600 courses taught by industry experts.
Scale and performance
From the course: Advanced Data Processing: Batch, Real-Time, and Cloud Architectures for AI
Scale and performance
- [Presenter] One of the key goals for an architecture is to be able to scale up, to meet the processing demands for the use case. When it comes to ML, scale and performance plays a key role in meeting the customer and business expectations. It is hence, an important focus area for ML architectures. What are the key enablers for scale and performance when it comes to ML? Let's briefly touch upon these enablers in this video. We will discuss more about them in the later chapters. Feature engineering requires data processing pipelines that can handle the incoming data volumes within expected latency thresholds. Feature engineering plays a key role in both model training and inference, so ensuring that these pipelines can scale is a critical responsibility for an AI architect. Next, comes the model architecture. The model architecture determines the kind of resources needed. It also determines the scale of inputs that can be processed. Choosing the right model architecture is critical to…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.