FDE-102
Forward Deployed Engineer Fast Track
The fast track through the core of the Agent Factory — for those who'd rather sprint than stroll. Command general agents, engineer the discipline that makes them dependable, and enter the market as a builder. Newer to this, or want a steady pace? Start at AI-101. Ready to invest the extra hours? Take the challenge.
Mode
Live Online Classes
Zia Tutor AI
Zia Tutor AI is a personal learning agent built as a digital twin of Zia Khan. It teaches the Agent Factory curriculum, remembers where you stopped, continues across sessions and weeks, teaches in the sequence appropriate for you, and checks whether you understood the previous concept before moving on.
Its mission is to help you gain the expertise to build AI Workers and Digital FTEs, and to become a Forward Deployed Engineer.
Learn MoreDetails
The fast track — for those who would rather sprint than stroll. The product isn't a website or a certificate; it's the ability to command AI agents that do real work, and the engineering discipline that makes them dependable. That discipline is the spine: spec-driven development, so the agent builds from a contract, not a vibe; loop engineering, so it plans, acts, and checks its own work; harness engineering, to make it reliable; trusting the checker, to verify with evals instead of reading every line; and leaving the laptop, to move it into a runtime that keeps working when you walk away. You run it across general agents — and run an open model locally, seeing that an agent is a harness plus a swappable brain. You solve real problems with the seven principles of problem solving, then ship what an AI worker is made of: a connector-native app any AI can use, agent identity, and a searchable System of Record on Postgres. You leave knowing where you fit in the market — the new roles and the FDE–Agent Factory model. No prior programming required: you'll learn to read, test, and verify AI-generated code, even when the agent writes most of it. The pace is demanding and weekly delivery is non-negotiable — a real challenge that will cost you extra hours. Newer to this, or want a steady climb? Start at AI-101 and follow the five-course path to Certified Agentic AI Architect at a normal pace. Ready to invest the hours? Take the challenge — a standalone accelerator with its own certificate, and a running start into the Professional Track.
What You'll Learn
Module 1Revision Sprint — Foundations & General Agents
A quick refresher to get everyone on the same page — prompting with context, teaching AI a task once with Skills and Connectors, and driving a general agent — so the real work starts on day one. Newer to this? AI-101 teaches these foundations in full.
Module 2Local AI Setup for Agentic Coding
Own the brain. Run an open model on your own machine and point a coding agent at it — private, offline, free. You see the idea under every AI tool made real: an agent is a harness plus a swappable brain, and the brain is just an address you can change.
Module 3Loop Engineering
Stop holding the tool turn by turn. Design an execution loop that plans, acts, checks its own work, and improves — a system that prompts the agent for you and calls you only at the gate. The leverage moves from the prompt you write to the loop you design.
Module 4Harness Engineering
A loop that runs while you sleep needs a layer that decides what the agent may do, what it knows, how its work is proven, and what happens when it breaks. That layer is the harness. Engineer it on purpose — permission walls, sandboxes, context, and automatic checks. Agent = model + harness.
Module 5Trusting the Checker
Every loop has a checker that says PASS or FAIL — but how do you know it's any good? Build evals: a folder of test cases drawn from real failures, grade the reviewer instead of trusting it once, and gate every change on the result. Verify agent output without reading every line.
Module 6Leaving the Laptop
A system proven on your laptop is trapped on one machine. Move it to a real runtime — a cloud schedule or a managed process — so it keeps working when you walk away, without losing the track record it earned. This is where your agent stops being a demo and starts being infrastructure.
Module 7Mode 1 — Problem-Solving
Solve it once. You drive the agent, the work gets done, you walk away.
Module 8Connector-Native Apps
Build a remote connector-native app — an MCP server whose customer is an AI, not a human. Because MCP is an open standard — a USB port for AI that every assistant fits — any AI on the internet can find your connector and use its tools, no custom wiring. This is how you give agents new hands: hosted, shareable capabilities they reach on their own.
Module 9AI Identity
Give an agent the right to act — safely. Human sign-in and agent access: how an AI proves who it's working for, what it may touch, and how it acts on your behalf without you handing over the keys to everything.
Module 10System of Record & RAG
Give your agent a memory it can search. Build a System of Record on Postgres with pgvector — a searchable knowledge and context store the agent queries to ground its work in your organization's truth, not its guesses.
Module 11The Agent Factory Ecosystem
Where you fit, and how you earn. The new AI roles — and the one the market can't find: the vendor-neutral Forward-Deployed Engineer. Then the FDE–Agent Factory model, a five-layer platform where graduates build and earn on top: Systems of Record for clients, manufacturing AI Workers, and domain startups they own. Postings up 800%+, no border, and Digital FTEs that become recurring revenue.
Course Outcomes
Command general agents to do real work, coding and non-coding alike
Run an open model locally — and see every agent as a harness plus a swappable brain
Engineer agents you can trust: specs, self-checking loops, and dependable harnesses
Verify with evals and checkers, not by reading every line
Ship your agent off the laptop into a runtime that keeps running without you
Solve real problems with method — and spot the ones worth turning into workers
Build what an AI worker is made of: a connector-native app, agent identity, and a searchable System of Record on Postgres
Know where you fit and how you earn: the FDE–Agent Factory model, and Digital FTEs as recurring revenue
Prerequisites
There are no pre-requisites for this course.