Choose this when
You have an AI app or workflow that needs reliability, cost control, latency, security, and observability.
All learning paths
AI learning path
Production AI engineering
Move from demos to AI systems that can be monitored, debugged, paid for, and improved.
- Best for
- Engineering teams
- Level
- Intermediate to advanced
- Time
- 16-30 hours
You should be able to
You can plan a production stack for prompts, models, retrieval, evals, tracing, deployment, and operations.
Checkpoint
Move on when you can answer: what changed, what broke, who noticed, and how will we prevent it?
Do
Learning sequence
This is the route through the topic. Watch and open the material inside the step where it is used.
Step 1
Instrument the system
Capture inputs, outputs, traces, costs, latency, user feedback, and model versions.
- Tracing
- Cost
- Latency
Watch here
Full Stack Deep Learning lecture
Full Stack Deep Learning
System-level view of shipping ML and AI products.
Open here
AI Engineering
Book · Chip Huyen · Intermediate to advanced
You are moving from demos to production systems.
Open resourceStep 2
Control releases
Use evals, canaries, fallbacks, and rollback plans for prompt and model changes.
- Canaries
- Fallbacks
- Rollbacks
Watch here
MLOps community production AI
MLOps Community
Use this for deployment, operations, and production system concerns.
Open here
Full Stack Deep Learning Lectures
Course videos · Full Stack Deep Learning · Intermediate to advanced
You want the whole lifecycle of ML and AI product development.
Open resourceStep 3
Operate the loop
Feed production failures back into data, prompts, retrieval, and product decisions.
- Feedback
- Datasets
- Iteration
Watch here
AI Engineering with Chip Huyen
Chip Huyen
Connects evals, systems, and production engineering judgment.
Open here
Phoenix by Arize
Open source tool and docs · Arize AI · Intermediate
You need to trace, inspect, and evaluate LLM app behavior.
Open resourceLangfuse Docs
Docs and cookbooks · Langfuse · Intermediate
You need production LLM tracing, scoring, and prompt operations.
Open resourcePractice task
Add tracing and a release checklist to one AI workflow, then use a real failure to improve it.
Reference
All resources in this path
Step 1
AI Engineering
Book · Chip Huyen · Intermediate to advanced
You are moving from demos to production systems.
Step 2
Full Stack Deep Learning Lectures
Course videos · Full Stack Deep Learning · Intermediate to advanced
You want the whole lifecycle of ML and AI product development.
Step 3
Phoenix by Arize
Open source tool and docs · Arize AI · Intermediate
You need to trace, inspect, and evaluate LLM app behavior.
Step 3
Langfuse Docs
Docs and cookbooks · Langfuse · Intermediate
You need production LLM tracing, scoring, and prompt operations.
Educators to follow
Chip Huyen
Intermediate to advanced
Use the book page and related essays as a production engineering path.
View educator
Eugene Yan
Intermediate to advanced
Read the applied ML and LLM systems posts.
View educator
Hamel Husain
Intermediate to advanced
Read the evals guide and build a small test set for your own app.
View educator
Lenny Rachitsky
Beginner to intermediate
Browse the How I AI interviews and copy the workflows that match your role.
View educator
Swyx
Intermediate to advanced
Watch AI Engineer talks for production patterns and tool choices.
View educator
Andrew Ng
Beginner to advanced
Start with ChatGPT Prompt Engineering for Developers, then pick a RAG or agents course.
View educator