</>SkillRecallRECALL WHAT YOU KNOW
Explore Library
AI FundamentalsAI Infrastructure and MLOps

10 items

1

Dynamic Quantization Before Inference

Code Quiz
2

Model Packaging Dependencies and requirements

Code Quiz
3

CI/CD Pipeline Stage Ordering for ML

Code Quiz
4

Canary and Shadow Deployment Routing

Code Quiz
5

Containerizing a Model with Dockerfile

Code Quiz
6

Data Drift Detection Computation

Code Quiz
7

Model Performance Monitoring in Production

Code Quiz
8

Concept Drift Monitoring Logic

Code Quiz
9

Batch vs Online Inference Pattern Selection

Code Quiz
10

GPU/TPU Acceleration and Resource Management

Slides / Video
Code QuizAdvanced

Model Performance Monitoring in Production

Fix the sliding-window monitor that discards the wrong sample.

Codepython
window = []

def monitor(prediction, label, size=100):
    window.append(prediction == label)
    if len(window) > size:
        window.pop()  # keep only the most recent `size` results
    return sum(window) / len(window)

What is the bug in this rolling accuracy monitor?