01
Foundations & Core Concepts
LLMs vs traditional development, frameworks, and tools.
›
- LLMs vs. Traditional Programming: How building with LLMs differs from and compares to standard software development.
- Industry Frameworks & Ecosystem: Understanding agents, MCP, LangGraph, and orchestration tools.
- How Tools Like Claude Code Work: Deconstructing CLI agents and simple execution loops under the hood.
02
Design Patterns & Architecture
Agents, RAG pipelines, and programmatic prompting.
›
- Agent Design Patterns: Implementing ReAct (Reason-Action), Plan-Execute, and custom deterministic flows.
- Building & Integrating RAG: Setting up a RAG pipeline and wiring it into your application flow.
- Programming (DSPy.ai & Jev Typesafe.ai) vs. Prompting: Program LLMs rather than relying purely on text prompts, thinking more as Input & Output.
03
Debugging, Costs & Observability
Guardrails, failure analysis, and token optimization.
›
- Production Guardrails & Safety: Moderation, Human-in-the-Loop (HITL), loop limits, and runtime boundaries.
- Debugging & Inspecting Decisions: Understanding why models fail, hallucinate, or deliver inconsistent results.
- Cost & Token Optimization: Designing your execution flow to cut token waste.
A realistic production AI scenario
The small cohort means we can spend time on your questions, architecture, and code.
Led by an engineer who builds these systems.
I'm Ahmad Baraka, an entrepreneur, software developer and engineering manager with 14+ years of experience building software and leading engineering teams.
I previously managed Mobile & Flutter engineering at Upwork, working with teams whose products serve millions of users.
Today I work primarily on AI and LLM engineering, including custom agent workflows, multi-repository retrieval systems, execution architectures and token-cost optimization.
LinkedIn profile →
Endorsed by engineering leaders now at Meta & Airbnb.
Recommendations from engineering directors and senior leads I worked alongside at Upwork who now lead teams at top global tech companies.
Former Engineering Manager leading multi-platform mobile engineering teams.
"I had the pleasure of working with Ahmad at Upwork, where he managed one of my teams, the Flutter Platform team. Ahmad consistently demonstrated strong leadership, proactive ownership, and sound judgment, he was effective at setting direction, aligning stakeholders, and keeping execution moving without sacrificing quality. It was particularly impressive to see him drive an innovative platform effort while maintaining a high bar for reliability and developer experience. I would gladly work with Ahmad again and recommend him for engineering leadership roles."
"Ahmad was on my team when I was his Tech Lead at Upwork. He stood out as someone who thought well beyond the code in front of him. He approached his work like an owner, not just a builder — he understood the bigger picture of what the business needed and regularly brought ideas on how to improve our processes, cut waste, and move faster. What made that even more valuable was that he wasn't just thinking about his own work — he was thinking about how the team as a whole could deliver better together. That kind of thinking came naturally to him, and it often pushed the rest of us to think differently too. On the engineering side, he was someone I could rely on. When Ahmad took ownership of something, it got done. He was also one of those people who made collaboration easy — calm, professional, and consistently solid to work with. Ahmad has had an entrepreneurial mindset for as long as I've known him, and I'm excited to see what he’s building next."
Built for engineers already building things.
Investment options.
-
Individual Cohort Seat Full 3-week cohort access$300 or 2 × $150
-
Private 1:1 Intensive Bring your own project & custom system emphasis$850
-
Corporate Workshop 1-day team workshop (Dubai)$3,000
Need team seats or a custom arrangement? Reach out via email to discuss multi-seat options.
Everything around the sessions.
-
6 Live Sessions
Interactive engineering sessions (~1.5 hrs each, twice per week).
-
Engineering Starter Kit
- Claude Code log parser to analyze step-by-step reasoning, cost, and flow.
- Agents setup flow to start building immediately.
-
30 Days Async Support
Private channel for architecture questions, code review, and guidance.
Build production AI systems.
If you're building with AI and want to understand what's happening beneath abstractions, let's talk.