PAYMENT METHOD
Price shown and payment processed in CFA francs. FedaPay transaction fees are added to the displayed amount, shown before you confirm payment.
See the course in action
Payment provider fees (Mobile Money, about 4%) are added at checkout.
By purchasing, you agree to our Terms of Sale.
PAYMENT METHOD
Price shown and payment processed in CFA francs. FedaPay transaction fees are added to the displayed amount, shown before you confirm payment.
- All content hours, with no expiry
- CodeLab in the browser, nothing to install
- AI tutor available at any hour
- AI-graded exercises
- Final exam and verifiable certificate + an optional field challenge for the Field Practice Validated distinction
Final exam: 2 attempts, 70% pass threshold. After a second failed attempt, a deferral period applies before retrying, with a €10 retake fee — see the Terms of Service.
Lifetime access. No deadline: finish in 10 weeks or in ten months.
LLM Engineering & Generative AI Applications
Build end-to-end LLM solutions: advanced RAG, AI agents, and technical security
- Level
- Advanced
- Duration
- 10 weeks
- Total hours
- 50 h
- Price
- 245 000 FCFA
The program, module by module
Module 1 — LLM Architecture
Understanding what happens between a prompt and its response — enough theory for good engineering decisions, no more.
Module 2 — The OpenAI API in Depth
Completion endpoints, function calling, structured outputs, streaming, and cost — building a complete Python client.
Module 3 — The Claude API in Depth
System prompts, tool use, extended thinking — and knowing when to prefer Claude over OpenAI.
Module 4 — Gemini and DeepSeek APIs
Gemini's native multimodality, DeepSeek's cost/performance ratio — building a multi-provider abstraction layer.
Module 5 — Advanced RAG
Intelligent chunking, re-ranking, hybrid RAG — improving a basic RAG and measuring the precision gain.
Module 6 — Fine-Tuning
Preparing a dataset, launching a fine-tune, evaluating the result against the base model — and knowing when to skip it.
Module 7 — Vector Databases
Choosing and sizing a vector database in production — migrating from in-memory storage to a managed database.
Module 8 — AI Agents and Rapid Prototyping
Advanced function calling, reasoning loops, multi-tool orchestration — demonstrable via Streamlit or Gradio.
Module 9 — Technical Security for LLM Applications
Prompt injection, jailbreak, context data leaks — compromising a vulnerable agent then fixing the flaws found.
Module 10 — Final Project and Certification
CertificationAssembling RAG, agents, prototyping, and security into an application connected to internal data, on a context of your choice — then building and certifying your final project.
How long will it take me?
3 h/wk ≈ 17 wks
6 h/wk ≈ 9 wks
12 h/wk ≈ 5 wks
20 h/wk ≈ 3 wks
Estimates, not deadlines. Pause for a month and pick up where you left off.
CERTIFICATE VERIFIABLE ONLINE
What you deliver at the end
Assembling RAG, agents, prototyping, and security into an application connected to internal data, on a context of your choice — then building and certifying your final project.
See a sample certificateHow certification works
Three levels, from simplest to most demanding. Each one only claims what it actually verifies.
Attendance certificate
Earned by: Modules followed
Proves: You followed the track — no graded assessment.
Certificate of achievement
Earned by: Full track completed on the platform, final project included, and final exam passed
Proves: Your knowledge and your final project, evaluated by the platform.
Field Practice Validated distinction
Earned by: A field challenge completed after the exam, in a real situation, with evidence
Proves: A CoursTech mentor reviewed and validated a genuine real-world application.
The distinction is added to your certificate of achievement — it never replaces it: nothing off-platform ever blocks the certificate itself.
See a sample certificate with the distinction →