Lyzr FDE Collective / The course

The world's most comprehensive FDE training program

AI Forward Deployed Engineer.

Name: The AI FDE Course
Price: 999 USD

Ten modules from a model call to a system running in a customer’s estate.

Enroll, $999 Explore all 100 competencies

Scope/Build/Harden/Deploy/Own

Engineers working through a production agent build

Every competency is taught on live systems and assessed on shipped work.

100

competencies mapped end to end

modules, each a production gate

50+

tools and platforms used hands-on

New here? Start with the Collective.

The path

Five stages. Ten modules. One deployment.

01
Scope
Find the workflow, qualify the value, agree what done means.
10Discovery, Delivery & Ownership
02
Build
Model behaviour, retrieval, architecture and integrations.
01Agent Foundations 02Knowledge & Retrieval 03Agent Architecture 04Multi-Agent & Integration
03
Harden
Evals, safety, identity and governance the review will test.
05Evaluation & Quality 06Safety & Security 07Data & Governance
04
Deploy
Promote into the customer estate and run it.
08Deployment & Operations
05
Own
Prove the economics, hand over, earn the expansion.
09Enterprise Systems & Economics

The complete curriculum

Ten gates between a model call and a system the customer trusts.

Open a module for its ten competencies and the tools you work with.

01
LLM behaviour, tokens & context windows
Lets you explain why an agent failed instead of guessing in front of the customer.
02
Model landscape & selection framework
Customers ask 'why this model' in every architecture review you will sit in.
03
Reasoning models & thinking budgets
Reasoning is a cost lever; leaving it on by default blows the budget at scale.
04
Reasoning & planning patterns (ReAct, CoT, Reflexion)
The shared vocabulary behind every framework and platform you will be handed.
05
System prompts & enterprise persona design
Tone, scope and refusal behaviour become contractual terms in regulated rollouts.
06
Structured outputs, schema enforcement & function calling
Downstream enterprise systems need typed, validated JSON, never prose.
07
Context engineering: budgeting, compaction & pruning
The single biggest driver of both cost and quality regressions in production.
08
Grounding, citations & source attribution
Enterprise answers must be traceable to a source document or they get rejected.
09
Prompt optimisation, versioning & rollout control
Prompts are code; unversioned changes silently break live customer workflows.
10
Multilingual & regional model choices
Global rollouts fail on language and script coverage long before they fail on accuracy.

Why this module matters: Every failure downstream traces back to a model or context decision made here.

How it is delivered

How it runs.

Format

Self-paced, cohort-supported

Work through the modules on your own schedule, with guided sessions and office hours running alongside.

Pace

10-12 weeks typical

Roughly 5 hours a week for working professionals. Access does not expire, so you can go faster or slower.

Practice

Build in every phase

Each phase ends in a hands-on build against enterprise-shaped constraints, not a quiz.

Capstone

Graded, end to end

One full engagement simulation, scope, build, ship, hand over, reviewed against the certification rubric.

The working stack

Fifty tools. Each one placed where it matters.

Five tools mapped to every module, the capability and where it sits in the system.

Agent Foundations

Knowledge & Retrieval

Agent Architecture

Multi-Agent & Integration

Evaluation & Quality

Safety & Security

Data & Governance

Deployment & Operations

Enterprise Systems & Economics

Discovery, Delivery & Ownership

A graded capstone certificate reviewed at a desk

Certification

Assessed on shipped work, not lesson completion.

The capstone follows the full arc, scope, build, harden, deploy, own, and is reviewed against the evidence an enterprise expects before go-live.

Level 1

Associate

You can execute the core patterns with guidance and explain the decisions behind them.

Level 2

Certified

You can deliver the full pattern independently in a production context.

Level 3

Lead

You can design for a novel enterprise problem and mentor a delivery team through it.

Readiness

Is this for you.

This is for you if

Comfortable with Python and APIs
Has called an LLM API and built a prompt-based app
Understands RAG and embeddings conceptually
Has used git in a real project

Not yet, if

Has never called an API
Does not yet know basic Python
Has not worked with version control
Looking for an introduction to programming rather than a production track

If that is where you are today, a fundamentals track is the better next step. The Academy will still be here.

Self-check

Ten questions. No account, nothing stored.

You write Python comfortably without a tutorial open.
You have called an LLM API directly, not only through a chat UI.
You have shipped something that uses a prompt in production or in a real pilot.
You can explain what an embedding is and when retrieval beats fine-tuning.
You have built or debugged a RAG pipeline.
You use git on real projects with other people.
You have integrated a third-party API with authentication and error handling.
You have worked inside an enterprise environment with security review.
You have owned a system after it went live, not only during the build.
You want a credential tied to production work, not lesson completions.

Answer the questions to see where you stand.

Apply to the Academy

Pricing

One price, the whole programme.

No per-module upsell and no certification fee on top. You pay once and keep access to the modules, the partner sessions and the Collective community.

AI FDE Certification

$999one-time, full programme access

All 10 modules and 100 mapped FDE competencies
Hands-on work across 50+ production tools and platforms
Graded capstone and the full agentathon simulation
Certification at Associate, Certified or Lead level
Access to the FDE Collective community and city chapters

Enroll now

Questions

Before you apply.

All 10 modules, 100 mapped competencies, hands-on work across 50+ tools and platforms, the graded capstone and certification. One-time payment, access does not expire.

AI Forward Deployed Engineer.

Five stages. Ten modules. One deployment.

Scope

Build

Harden

Deploy

Own

Ten gates between a model call and a system the customer trusts.

LLM behaviour, tokens & context windows

Model landscape & selection framework

Reasoning models & thinking budgets

Reasoning & planning patterns (ReAct, CoT, Reflexion)

System prompts & enterprise persona design

Structured outputs, schema enforcement & function calling

Context engineering: budgeting, compaction & pruning

Grounding, citations & source attribution

Prompt optimisation, versioning & rollout control

Multilingual & regional model choices