CS 203: Software Tools and Techniques for AI (STT-AI)

Aug–Nov 2026 | IIT Gandhinagar

Welcome

Welcome to the course website for CS 203: Software Tools and Techniques for AI (STT-AI). This course introduces practical tools and techniques for building reproducible, efficient, and deployable artificial intelligence systems—from data collection and labeling to model deployment and continuous integration.

Credits: L–T–P–C: 1.5–0–1.5–4

Prerequisites: None

Honour Code: View Policy

Course Overview

  1. Data collection and labeling: Collecting data using tools for logging and instrumenting sources such as websites; validating data; labeling data using tools such as Label Studio; inter-annotator agreement using methods such as Cohen's kappa; snowball sampling and active learning for cost-effective annotation; and data augmentation.
  2. Model and code reproducibility and versioning: Architecture and hyperparameter search using tools such as Weights & Biases and AutoML tools such as AutoGluon; model checkpointing using MLflow, TensorBoard, and Weights & Biases; and data versioning.
  3. Self-contained environments: Using tools such as Docker to manage released packages and working with cloud platforms such as Google Cloud Platform (GCP), Amazon Web Services (AWS), and Microsoft Azure.
  4. Run-time profiling and optimization: Measuring and optimizing memory footprint, GPU utilization, and other run-time performance characteristics.
  5. Continuous integration and continuous deployment: A/B testing; distribution and covariate shift; data version management and retraining to account for shifts; and redeployment.
  6. Deployment on constrained devices: Tools such as TensorFlow Lite and techniques such as model quantization.

Recent Announcements

Coming soon: Course announcements and materials will be posted here.