OneRange

Embeddings & Vector Databases

A deeper technical layer under RAG: choosing embedding models, operating vector databases, and tuning hybrid search. Learners benchmark retrieval quality and cost across configurations.

What this course covers

The course runs across 10 topics, each one a short adaptive session rather than a recorded lecture. The tutor explains the idea, works an example, checks understanding, and adjusts the next step to the answer given.

  • Selecting and Evaluating Embedding Models — Compare embedding models based on dimensionality, context limits, and domain suitability.
  • Vector Store Architectures and Deployment — Understand the architecture of vector databases and perform initial setup of a vector database instance.
  • Configuring Vector Indexes: HNSW vs. IVF — Configure and test different indexing algorithms (Flat, IVF, HNSW) to balance search latency and recall accuracy.
  • Distance Metrics and Similarity Search — Analyze the mathematical foundations and practical trade-offs of Cosine, L2, and Dot Product distance metrics.
  • Metadata Filtering and Indexing Strategies — Implement pre-filtering, post-filtering, and single-stage filtering to narrow down vector searches using metadata.
  • Implementing Sparse-Dense Hybrid Search — Combine dense vector embeddings with sparse keyword search (BM25) to improve retrieval relevance across diverse query types.
  • Result Merging with Reciprocal Rank Fusion (RRF) — Implement RRF and integrate cross-encoder re-ranking models to optimize final retrieval results.
  • Scaling Vector DBs: Sharding and Partitioning — Set up database partitioning, sharding, and replication strategies to handle millions of vectors efficiently.
  • Benchmarking Retrieval Quality and Latency — Write a benchmarking script to measure Recall@K, NDCG, and latency across different vector configurations.
  • Memory Optimization and Quantization — Apply Scalar Quantization (SQ) and Product Quantization (PQ) to reduce index size and optimize memory usage.

Skills you build

Results are measured against named skills in the OneRange taxonomy of more than 10,000 skills, so a manager sees proficiency per skill rather than a completion tick.

Who it is for

Builders, data engineers. The material is pitched at advanced level, and takes roughly 150 minutes at a typical pace. Because every session adapts, someone who already knows a topic moves through it quickly instead of sitting through an explanation they do not need.

It sits in the Category: RAG track of the Vero AI Catalog, and can be assigned to one person, a team, or the whole company.

How it is delivered and assessed

Delivery is conversational and interactive, including code lab, guided lab exercises. There is no video to sit through and no slide deck to click past.

Understanding is checked with a code assessment of 5 items, with a pass mark of 70%.

Administrators can copy this course into their own library and adapt it — edit the outline, change the duration, swap the assessment format, or ground it in internal documentation so answers cite the company's own source material.

Also in this track

  • RAG & Knowledge Systems — This course teaches builders to create and evaluate retrieval-augmented systems that ground AI in their data. It covers why RAG matters, embeddings and vector search, building a knowledge base, and evaluation and pitfalls. Learners finish able to build a RAG system that returns trustworthy answers.
  • Modern AI Stack for Builders — Master the integration layer of modern AI systems

Related

Frequently asked questions

How long does this course take?

It runs roughly 150 minutes at a typical pace. Because every session adapts, someone who already knows a topic moves through it quickly instead of sitting through an explanation they do not need.

How is it assigned to a team?

An administrator assigns it from the Vero dashboard to one person, a team, a department or the whole company, and sees progress and results per person and per skill.

Can we customise it with our own documents and terminology?

Yes. Administrators can copy this course into their own library and adapt it — edit the outline, change the duration, swap the assessment format, or ground it in internal documentation so answers cite the company's own source material.

Course code CAT-VDB.