BestsellerIntermediate

Applied AI: Build with Language Models

4.8 (1,284)
· 12,840 learners
Course preview: a laptop showing the extraction pipeline built in the project
4h 07m
13 lessons
English
Self-paced or cohort

About this course

Most courses show you what a model can do. This one makes you build the three features every product team is asked for — retrieval that answers the right question, output you can hand to other code, and an evaluation harness that catches regressions before your users do. Every module ends in something that runs; the reviewed project is a working extraction pipeline you can adapt.
Skills you practice
Retrieval-augmented generationStructured outputEvaluation harnessesPrompt designVector search

Syllabus

4 modules · 13 lessons
What a language model can and cannot promise
12 min
The working setup: SDK, keys, and a scratchpad
18 min
Tokens, temperature, and the cost dial
18 min
Chunking strategies that survive real documents
21 min
Vector search from first principles
19 min
Rerankers and the relevance ceiling
18 min
When the model should say it found nothing
18 min
Schemas as contracts, not suggestions
22 min
Validation, repair, and retry loops
21 min
Streaming partial results into a UI
21 min
Golden sets and where they come from
20 min
Scoring retrieval and generation apart
19 min
Regression gates in continuous integration
20 min

Your instructor

Gabriela FernándezApplied AI engineer, formerly search infrastructure
12,840 learners
4 courses
4.9 instructor rating
Gabriela spent nine years building retrieval systems at scale before language models landed on her desk. She has shipped the exact pipeline this course teaches — twice, in production, with the scars to prove which parts matter.

Reviews

How reviews work
4.8
Rated 4.8 out of 5
1,284 reviews
Content depth4.9
Content depth
Practical exercises4.8
Practical exercises
Instructor clarity4.9
Instructor clarity
Pacing4.6
Pacing
Materials and project4.7
Materials and project
Community support4.5
Community support
Drew YoungReviewed 2 weeks ago
Rated 5 out of 5
The evaluation module alone is worth the price. I walked in with a half-working RAG demo and walked out with a harness that caught two regressions in our first sprint after launch.
Jihoo SongReviewed a month ago
Rated 5 out of 5
Gabriela explains rerankers better than the papers do. The chunking lesson changed how our team splits documentation, and the project review was concrete, not rubber-stamp.
Ana ThomasReviewed 2 months ago
Rated 4 out of 5
Dense but fair. I wanted more on fine-tuning and got none — the course is honest about scoping to what ships. The structured-output lessons replaced three fragile regex pipelines of mine.
Pablo MoralesReviewed 3 months ago
Rated 5 out of 5
I have taken four applied-AI courses this year and finished one. This is the one. Every module ends in something that runs, exactly as promised, and the finish-date picker kept me honest about pacing.

Common questions

No. Every exercise runs on a free-tier API key, and the project review accepts either the free tier or a local model. The course states where results differ between them.
Access is lifetime, including every future revision of the lessons. When the tooling shifts — and it will — the affected lessons are re-recorded rather than patched with errata.
Every purchase carries a 30-day money-back guarantee, no questions and no forms beyond one click. Certificates already issued are yours to keep.
Yes — a graded certificate issued when the reviewed project passes. It lists the pipeline you built, not just attendance.
$89$129Sale
One-time purchase · yours to keep
30-day money-back guarantee
Start
Weekly pace
At 7 h/week — 6 weeks · finish around Apr 13
This course includes
Lifetime access to 13 lessons and all future revisions
The reviewed project: a working extraction pipeline
A graded certificate of completion
Offline downloads for every lesson