Reproducible training pipelines in MLOps - Time & Space Complexity

Time Complexity: Reproducible training pipelines

O(n)

Understanding Time Complexity

When building reproducible training pipelines, it's important to know how the time to run the pipeline changes as the data or steps grow.

We want to understand how the pipeline's execution time scales with input size.

Scenario Under Consideration

Analyze the time complexity of the following code snippet.

for batch in data_batches:
    preprocess(batch)
    train_model(batch)
    validate_model(batch)
    save_checkpoint()

This code runs a training pipeline on batches of data, processing each batch through preprocessing, training, validation, and saving checkpoints.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: Loop over each data batch running all pipeline steps.
• How many times: Once per batch, so the number of batches determines repetitions.

How Execution Grows With Input

As the number of batches increases, the total time grows roughly in direct proportion.

Input Size (n)	Approx. Operations
10	10 times the pipeline steps
100	100 times the pipeline steps
1000	1000 times the pipeline steps

Pattern observation: Doubling the number of batches roughly doubles the total work.

Final Time Complexity

Time Complexity: O(n)

This means the total time grows linearly with the number of data batches processed.

Common Mistake

[X] Wrong: "The pipeline time stays the same no matter how many batches there are."

[OK] Correct: Each batch requires running all steps, so more batches mean more total work and longer time.

Interview Connect

Understanding how pipeline time scales helps you design efficient workflows and explain your approach clearly in real projects or interviews.

Self-Check

"What if we parallelize processing batches instead of running them one by one? How would the time complexity change?"