AI 용어집
인공지능 완전 사전
Model and data versioning
Version tracking systems for datasets, features and ML models ensuring reproducibility and traceability
CI/CD Pipelines for Machine Learning
Automation of build, test and deployment of ML models with adapted continuous integration and continuous deployment
Model monitoring and observability
Real-time monitoring of performance, data drifts and abnormal behaviors of models in production
ML infrastructure management
Orchestration of computational resources for ML model training, deployment, and inference
Feature Engineering and Feature Stores
Centralization and management of features to ensure consistency between training and production
A/B Testing and experimentation
Comparative testing frameworks to evaluate and progressively deploy new models
MLOps Platforms and tools
Integrated solutions to orchestrate the entire ML lifecycle from experimentation to production
Model Deployment
Deployment strategies and patterns including batch, real-time, serverless and edge computing
Governance and ethics in production
Compliance frameworks, auditability, and bias management for AI systems in production
AutoML and automation
Tools for automating feature engineering, model selection, and hyperparameter optimization
Edge MLOps
Deployment and management of ML models on edge devices with resource and connectivity constraints
Model lifecycle management
Complete orchestration of model lifecycle from creation to retirement through maintenance