deployed_code

MLOps & LLMOps

From experiment tracking to production serving — the engineering practices, tools, and platforms that take ML models from notebook to production
Co-Created by Kiran Shirol and Claude
Topics Experiment Tracking CI/CD for ML Model Serving LLMOps Monitoring Platforms
home Learning Portal play_arrow Start Learning summarize Key Insights dictionary Glossary 10 chapters · 3 sections
Section 1

Foundations — The MLOps Lifecycle

Why ML in production is hard, and the tools to tame the chaos.
1
warning
Why MLOps Matters
Technical debt in ML, the “Hidden Technical Debt” paper, the MLOps lifecycle, and maturity levels.
arrow_forward Learn
2
science
Experiment Tracking
MLflow, Weights & Biases, experiment logging, hyperparameter tracking, and reproducibility.
arrow_forward Learn
3
inventory_2
Model Registry & Versioning
Model artifacts, versioning strategies, staging/production promotion, and DVC for data.
arrow_forward Learn
4
database
Data Pipelines & Feature Stores
ETL for ML, Feast, Tecton, feature engineering at scale, and data quality checks.
arrow_forward Learn
Section 2

Deployment — CI/CD, Serving & LLMOps

Getting models into production and keeping them there — including LLM-specific operations.
5
sync
CI/CD for Machine Learning
Testing ML code, model validation gates, GitHub Actions for ML, CML, and automated retraining.
arrow_forward Learn
6
dns
Model Serving & Inference
TorchServe, Triton, vLLM, ONNX Runtime, batching strategies, and latency vs. throughput.
arrow_forward Learn
7
route
LLMOps: Gateways & Routing
LLM gateways (LiteLLM, Portkey), model routing, fallbacks, rate limiting, and cost tracking.
arrow_forward Learn
8
edit_note
LLMOps: Prompt Management & Evaluation
Prompt versioning, A/B testing prompts, LLM evaluation in CI, and guardrails integration.
arrow_forward Learn
Section 3

Operations — Monitoring & the Full Stack

Keeping models healthy in production and choosing the right platform.
9
monitoring
Monitoring & Drift Detection
Data drift, concept drift, model performance decay, Evidently AI, Arize, and alerting strategies.
arrow_forward Learn
10
stacks
The MLOps Stack
End-to-end platforms (Kubeflow, SageMaker, Vertex AI, Azure ML), choosing your stack, and build vs. buy.
arrow_forward Learn
explore

Explore Related Courses

memory
AI Infrastructure
GPUs, Clusters & Networking
checklist
LLM Evaluation
Benchmarks, Metrics & Testing
smartphone
Small Models & Local AI
Quantization & Edge Deployment