The Big Picture: Where Model Training Fits in AI Development


By Sam
This description is based on a conversation with 'Sam', which took place over several days in January 2025

The model training phase is just one part of a much larger AI development pipeline:
  • Pre-Processing: Cleaning, structuring, and preparing training data.
  • Post-Preprocessing: Feature selection, embedding generation, and final dataset refinement.
  • Model Training: The learning phase, where the model improves through cycles of optimization.
  • Fine-Tuning & Evaluation: Adjusting for real-world performance using human feedback and task-specific data.


Each step in model training brings AI closer to accuracy, efficiency, and real-world usability. With models growing in complexity, efficient training techniques and compute optimizations are more critical than ever.