AI Forward Deployed Engineer (AI FDE) Professional Program

A 12-week, live online weekend program that takes a working developer to a job-ready AI Forward Deployed Engineer: someone who can scope a client's problem, build an agentic AI solution, deploy it, and prove its value.

All courses / Artificial intelligence

Course at a glance

Mode: 100% online, live instructor-led, recordings provided
Duration: 3 months (12 weeks)
Class days: Saturday and Sunday
Class time: 7:30 pm to 9:00 pm (Pakistan time)
Live sessions: 24 sessions of 90 minutes, 36 live hours
Self-study and labs: 5 to 6 hours per week (about 65 hours)
Total effort: About 100 hours
Projects: 2 guided projects and 1 client-style capstone
Outcome: Deployed portfolio of 3 projects, AiPearlz certificate, interview and freelancing readiness

Each weekend follows one rhythm. Saturday teaches the concept with a live demo. Sunday is a hands-on build lab on the same topic. A weekday assignment carries the build forward.

What makes this program different

A pace built for working professionals: two evening classes each weekend over 12 weeks.
Sunday build labs with an instructor present, so nobody is stuck alone on setup or debugging.
Mock client engagements: discovery call, scoping document, demo and handover, all assessed.
Local career track: portfolio review, remote job applications and freelancing.

By the end, you can

Explain how LLMs work and choose a model on quality, speed, cost and privacy.
Build RAG systems and tool-using agents, with and without frameworks.
Design multi-agent architectures with memory, durable state and human approval.
Measure quality with evals and observability, and use the results to improve a system.
Secure an agent against prompt injection and data leakage with guardrails and gateways.
Deploy an AI application to the cloud with Docker and CI/CD, and monitor cost and latency.
Run a client engagement: discovery, scoping, success metrics, demo, handover.
Present a 3-project portfolio and pass an FDE-style interview.

Who should join

Software developers and fresh CS/IT graduates moving into AI engineering
Data analysts, data engineers and ML practitioners who want to ship LLM products
Solution architects, technical consultants and freelancers selling AI projects to clients

Prerequisites

Working Python: functions, classes, packages, calling a REST API
Basic Git and command line
A laptop with 8 GB RAM or more and a stable internet connection
No prior machine learning or LLM experience needed

Program structure

3 phases, each ending in a demo you must pass to move on.

1
AI Engineer
Weeks 1 to 4
LLMs, prompting, RAG, first agents
You build: Project 1: company knowledge assistant with citations
To pass: Live demo and code review, end of Week 4
2
Agentic AI Architect
Weeks 5 to 8
Frameworks, evals, multi-agent design, AI-assisted development
You build: Project 2: multi-agent workflow automation with evals
To pass: Live demo and architecture review, end of Week 8
3
Forward Deployed Engineer
Weeks 9 to 12
Security, deployment, client engagement, career
You build: Capstone: client-style solution, deployed and handed over
To pass: Demo day in front of a panel, Week 12

Week-by-week curriculum

Saturday is concept plus live demo. Sunday is a guided build lab. Both run 7:30 pm to 9:00 pm.

Month 1: Become an AI Engineer

Week 1
Saturday: concept and demo
Orientation. The FDE role, where it came from, job market and roadmap. How LLMs work: tokens, context windows, training, the major labs and models.
Sunday: build lab
Environment setup: Python, uv, Git, VS Code or Cursor, API keys, OpenRouter free models. First LLM API calls.
Weekday assignment
Call three different models with the same prompt and compare output, speed and cost.
Week 2
Saturday: concept and demo
Prompt engineering, system prompts, structured outputs. Hallucination and bias: why they happen and how to reduce them.
Sunday: build lab
Build a chat application with FastAPI and a Gradio front end. Streaming and conversation memory.
Weekday assignment
Add structured JSON output and input validation to the chat app.
Week 3
Saturday: concept and demo
Model selection: frontier, open-source and local models on quality, speed, token cost and privacy. Embeddings, vector databases and RAG.
Sunday: build lab
Build a RAG pipeline: chunking, embeddings, Chroma or Qdrant, retrieval, answers with citations. Run a local model with Ollama.
Weekday assignment
Start Project 1 on a document set of your choice.
Week 4
Saturday: concept and demo
Agents from first principles: the agent loop, tool calling, memory. Model Context Protocol (MCP).
Sunday: build lab
Build an agent without a framework, then connect it to tools through MCP. Project 1 demos.
Weekday assignment
Polish Project 1, write the README, publish to GitHub.

Month 2: Become an Agentic AI Architect

Week 5
Saturday: concept and demo
Agent frameworks and harness engineering: LangChain and LangGraph, Claude Agent SDK, OpenAI Agents SDK. When to use a framework and when not to.
Sunday: build lab
Rebuild the Week 4 agent in LangGraph. Add planning and sub-agents.
Weekday assignment
Compare the framework and no-framework versions in a short write-up.
Week 6
Saturday: concept and demo
Evals and observability: golden datasets, LLM-as-judge, tracing, cost and latency tracking with Langfuse.
Sunday: build lab
Instrument Project 1 with Langfuse. Build an eval set and improve the score with evidence.
Weekday assignment
Reach an agreed eval target on Project 1 and report what changed.
Week 7
Saturday: concept and demo
Architecture patterns: agentic RAG, durable state, human-in-the-loop approval, multi-agent orchestration and swarms. Fine-tuning versus RAG versus prompting.
Sunday: build lab
Build a multi-agent workflow with persistent state and a human approval step.
Weekday assignment
Start Project 2: architecture diagram and eval plan.
Week 8
Saturday: concept and demo
AI-assisted development: Claude Code and the Agentic Development Lifecycle. Specs, reviews, tests and staying accountable for AI-written code. Data fundamentals and annotation for evals.
Sunday: build lab
Build sprint on Project 2 with coding agents. Project 2 demos and architecture review.
Weekday assignment
Finish Project 2 and publish it.

Month 3: Become a Forward Deployed Engineer

Week 9
Saturday: concept and demo
Security and responsible AI: prompt injection, the lethal trifecta, adversarial attacks, guardrails, policies, AI gateways, data privacy and governance.
Sunday: build lab
Red-team lab: attack a classmate's agent, then harden your own. Capstone briefs issued.
Weekday assignment
Security report on Project 2. Choose a capstone brief.
Week 10
Saturday: concept and demo
Client engagement, part 1: discovery calls, mapping the workflow, scoping, success metrics, ROI, writing a proposal. Choosing between AI, automation and ordinary software.
Sunday: build lab
Role-play: discovery call with a mock client. Draft the scope document.
Weekday assignment
Submit capstone scope document and architecture.
Week 11
Saturday: concept and demo
Production deployment: Docker, CI/CD with GitHub Actions, cloud hosting, secrets, monitoring, cost control. Overview of managed agent platforms on AWS, Azure and Google Cloud.
Sunday: build lab
Deploy the capstone to the cloud with a pipeline and live monitoring.
Weekday assignment
Capstone build. Record a 5-minute demo video.
Week 12
Saturday: concept and demo
Client engagement, part 2: demos, stakeholder updates, handover and documentation. Career session: portfolio, GitHub, LinkedIn, FDE interviews, remote jobs and freelancing.
Sunday: build lab
Capstone demo day before a panel. Feedback, results and certificates.
Weekday assignment
Graduate: final portfolio submission.

Projects and capstone

You leave with 3 public, deployed projects, each one harder and closer to real client work.

1
Company knowledge assistant due Week 4
RAG chatbot over a real document set (policies, manuals, product catalogue).
Must include: Citations, a simple web UI, README, GitHub repo
2
Multi-agent workflow automation due Week 8
Agents that complete a business process end to end, such as support ticket triage, invoice processing or lead research.
Must include: At least two agents, tool use, human approval step, eval set with scores, Langfuse traces
3
Capstone: client engagement due Week 12
A solution for a mock or real client brief, taken from discovery to handover.
Must include: Scope document, architecture diagram, deployed app, security review, eval report, demo video, handover document

Capstone options

Customer support agent for an e-commerce store, answering in English and Urdu
Admissions and student-query assistant for a school or institute
Document intake and summarisation agent for a law or accounting firm
Sales lead qualification and follow-up agent for a real estate agency
Appointment and patient-query assistant for a clinic

You may bring a brief from your own employer or a freelance client, with instructor approval.

Tool stack

The stack favours free tiers and open-source tools, so you pay little beyond the course fee.

Area	Tools taught	Cost to you
Language and setup	Python, uv, Git and GitHub, VS Code or CursorFree	Free
Models	OpenRouter free models, Ollama for local models, plus one frontier API (Claude, GPT or Gemini)Free to a few dollars of API credit	Free to a few dollars of API credit
Apps and APIs	FastAPI, Gradio or StreamlitFree	Free
RAG	Chroma or Qdrant, embedding modelsFree	Free
Agents	Plain Python agent loop, LangChain and LangGraph, Claude Agent SDK, OpenAI Agents SDK, MCPFree	Free
Evals and observability	LangfuseFree tier	Free tier
AI-assisted development	Claude Code, or a free coding agent as an alternativePaid subscription optional	Paid subscription optional
Deployment	Docker, GitHub Actions, one cloud free tier (AWS, Azure or Google Cloud), or Render or RailwayFree tier	Free tier
Security	Guardrail libraries, an AI gateway such as LiteLLMFree	Free

Assessment

Weekly assignments and quizzes: 20%
Project 1: 15%
Project 2: 20%
Capstone (build, documents, demo): 35%
Attendance and participation: 10%

Certificate requirements

Certificate of Completion: 60% overall, 75% attendance (live, or recording with quiz) and the capstone submitted.
Certificate with Distinction: 85% overall and a capstone rated production-ready by the panel.
Each certificate carries a verification link or QR code to the student's portfolio.

Your instructor

M. Iqbal Gill

Agentic AI Engineer and Lead AI Instructor

Agentic AI Engineer and Lead AI Instructor with 7+ years of experience delivering industry-focused, project-based technical training to 1,000+ students across national AI and web development programs. A specialist in Generative AI, multi-agent systems, RAG architectures and enterprise AI solutions, with deep full stack development experience across modern JavaScript, Python and cloud-native tooling.

Full profile LinkedIn

Upcoming batches

Classes run Saturday and Sunday, 7:30 pm to 9:00 pm, Pakistan time.

Start dates will be announced soon. Apply now and we will tell you as soon as they are set.