RAG and search engineer
Learn to build search and question-answering over private documents that returns the right answer and proves it with tests, working with AI coding tools from week 1, in 20 weeks.
Coming soon · planned, not written yet
- 20
- weeks planned
- ~40 h
- a week
- 5
- milestone projects
- 1
- project, grown every week
These are planned figures. The weeks are written after we check the tools against current job postings and official documentation, so the final route may differ.
Built for the AI era
How AI fits this roadmap
AI coding assistants now write much of the glue code around embeddings, vector stores and pipelines, so the valuable part of this job is deciding what to retrieve and proving it works. This roadmap teaches you to use an AI assistant every week and to check its output with tests and measured retrieval scores. Retrieval is also the content itself: you build, evaluate and secure the search layer that language-model applications and agents depend on.
The main project
AskDocs
A question-answering service over a real document collection, with search, citations and a test set, that you grow from a first script into a measured, deployed API.
What you will be able to do
- 01
Turn documents into searchable chunks with metadata, and explain why each chunking choice was made.
- 02
Build keyword, vector and hybrid search with Postgres and pgvector and with Qdrant, and compare them on the same questions.
- 03
Add reranking with a cross-encoder model and show with numbers whether it helped.
- 04
Write a test set of questions and expected sources, and score every change with retrieval metrics and a RAG evaluation tool such as Ragas.
- 05
Deploy the service as a FastAPI container with citations, logging and access control, and expose it to AI assistants through an MCP server.
01The route
Five months, one month at a time.
Each month ends in a milestone project, so you always have something to show for the work.
Month 1 · Weeks 1–4
Python, text and AI-assisted work
Python for data work, SQL, and how language models read text. You set up an AI coding assistant, learn to prompt it, read its diffs and test what it writes. You also build the first keyword search over documents.
MilestoneA command-line keyword search over a real document set, with tests and a log of AI suggestions you accepted and rejected
Month 2 · Weeks 5–8
Embeddings and vector search
Embedding models with Sentence Transformers, chunking, and nearest-neighbour search with pgvector. You learn what the index settings trade off and use AI to draft queries, then check them on real data.
MilestoneSemantic search over your documents in Postgres, with chunks, metadata and a first answer step that cites its sources
Month 3 · Weeks 9–12
Hybrid search, reranking and Qdrant
Combine keyword and vector results, add metadata filters and a cross-encoder reranker, and rebuild the same search in Qdrant. You compare the two stores on the same questions.
MilestoneA hybrid search with reranking on two vector stores, and a short written comparison
Month 4 · Weeks 13–16
Evaluation and AI quality
Build a test set of questions and expected sources, measure recall and rank quality, and score answers with Ragas. You learn where AI-written answers go wrong and how to catch it.
MilestoneAn evaluation report that scores every change to AskDocs, with a test that fails when quality drops
Month 5 · Weeks 17–20
Production, agents and capstone
Serve AskDocs with FastAPI in Docker, add logging, access control and cost tracking, expose it to AI assistants through an MCP server, then write the portfolio and prepare for interviews.
MilestoneA deployed, documented AskDocs API with an MCP server, an evaluation report and a portfolio write-up
03What is inside
Everything a week comes with.
Every week is built the same way, so once you have done one you know how all of them work.
- 01
One project, AskDocs
You grow one question-answering service from a script to a measured API, so every tool lands in something real.
- 02
A day-by-day plan
Each week has a Learn part and a Do part, with a checklist so you know when you are done. Most weeks include a task to do with an AI assistant and a step to check its work.
- 03
A project in every milestone
Five milestones, each ending in something you can show and link to.
- 04
Reference topics
Short topic pages with check questions you can come back to.
- 05
A job track
Interview practice, portfolio and resume help, and an application plan.
- 06
Progress in your Study plan
Add weeks to your Study plan, give them dates, and finished days are counted for you.
04Is it for you
Who this is, and is not, for.
Better to know before you start than six weeks in.
It is for you if
- You can write small programs in Python or a similar language and have used a database.
- You are a developer or data person who wants to build the retrieval side of AI applications and be strong with AI tools.
- You can give about 40 hours a week for five months.
It is not for you if
- You have never written code. Do a beginner course first.
- You want a job guarantee. Nobody can promise that, and this roadmap does not.
- You want to train large models from scratch. This roadmap uses existing models and focuses on retrieval.
- You want to paste AI output into production without checking it. This roadmap trains you to measure it.
?Questions
Before you start
- Is this roadmap ready?
- Not yet. This is the plan. The weeks are being written and checked against current job postings and official documentation. Leave your email and we will write to you when it opens.
- What will it cost?
- The price is not set. Roadmaps are bought once, with no subscription. We will announce it when this one opens.
- Will the plan change?
- It may. Before the weeks are written we check the tools against current job postings, so the lengths and months here are planned, not final.
- Do I need a paid AI model account?
- Probably a small one. The roadmap will use a hosted language model for answers and an AI coding assistant, and will say which parts work with free or local models once the weeks are written.
- Will it get me a job?
- It will give you skills and proof of work. It cannot promise a job on a fixed date, because hiring depends on the market, your location and your past work.
Not open yet
Leave your email and we will write once, when it opens.