Embeddings Engineering
Stop treating embeddings as a black box — understand, tune, and monitor them.
Embeddings underpin every RAG system, semantic search, and similarity feature — but most engineers treat the embedding model as an opaque API call. This course covers how embeddings work, reading the MTEB leaderboard for your task, fine-tuning for domain-specific retrieval (15–30% quality gains), chunking strategies (parent-child, structural), hybrid dense + sparse retrieval with RRF, embedding drift monitoring, and a domain-adapted pipeline capstone.
7h
Duration
8
Lessons
0
Learners
Course map
Lessons unlock as you complete the previous one. Your progress is saved on this device.
Lesson 1
What embeddings are and why they work
9m33 XPLesson 2
Choosing an embedding model: MTEB and task-specific evaluation
11m38 XPLesson 3
Fine-tuning embedding models for domain-specific retrieval
13m45 XPLesson 4
Chunking strategies for high-quality embeddings
10m35 XPLesson 5
Embedding drift and production monitoring
10m35 XPLesson 6
Hybrid retrieval: dense + sparse
10m35 XPLesson 7
Embedding evaluation: offline metrics and production quality signals
10m35 XPLesson 8
Capstone: build a domain-adapted embedding pipeline
18m60 XP
Take next
Courses that pair well after — or alongside — Embeddings Engineering.