01Foundations
Agentic Engineering
From AI that autocompletes to AI that acts. The shift, the loop, what it can and can’t do yet.
Recorded + Live Q&A
Agentic Engineering course
A 12-part course for engineers who want to build production software with Claude Code.
What this is
Most teams have AI tools installed and nothing has changed. The copilots autocomplete, the chats answer, and the work still goes at the same pace it did before — sometimes slower. The problem isn’t the model. It’s the workflow around it.
This course is the workflow. Twelve hands-on sessions that take you from a coding agent that feels novel to one that quietly does real work in the background. Context, planning, specs, skills, subagents, MCP — all the pieces that turn a chatbot into a shipping teammate.
The agent does the typing. You still own the architecture.
What we cover
- Claude Code
- Context Engineering
- Plan Mode
- CLAUDE.md
- Skills
- Subagents
- Hooks
- MCP
- OpenSpec
- GitHub Actions
- Conductor
- Codex Reviewer
The syllabus
Twelve parts. One workflow.
01 — 12
02Foundations
Claude Code 101
A terminal-native coding agent. Tools, edit modes, models, the agentic loop, slash commands.
03Foundations
Context Management
Curate what the agent sees. Fill the window well. Bad runs are usually a context problem.
04Workflow
The Core Workflow
Plan → Execute → Verify. A repeatable loop for production-grade work with a paper trail.
05Workflow
CLAUDE.md
The README for the agent. The one file every repo should have, and how to keep it healthy.
06Extensions
Skills
Reusable, named workflows the agent loads only when it needs them. Build once, reuse forever.
07Extensions
Subagents
Specialist coworkers with their own context window. Delegate the noisy work, keep the main clean.
08Extensions
Hooks
Lifecycle callbacks for the agent loop. Before-tool gates and after-tool notes — format, lint, block.
09Extensions
MCP Servers
The USB-C of AI tools. One protocol to graft GitHub, Postgres, Sentry, Figma onto Claude.
10Production
Spec-Driven Development
Agree on what before how. OpenSpec, delta specs, four verbs in one loop.
11Production
GitHub Integration
Claude inside your repo. Triggered by issues, PRs, and CI failures — not just prompts.
12Production
Parallel with Conductor
N Claudes in isolated git worktrees. Three features in one afternoon. One reviewer: you.
Who it’s for
For engineers who want to ship with the agent.
- You’ve tried Claude Code or Cursor and want to go from “it’s neat” to “it ships”.
- You write production code daily and want a real workflow, not a vibes-driven demo.
- You’re comfortable in a terminal and want to use the agent like a senior engineer would.
- You want a repeatable loop — plan, execute, verify — you can apply on Monday morning.
- You’d rather build with the agent than watch a slide deck about it.
Who teaches it
Mohammad Emran Hasan
Co-Founder & CEO of Klasio / FigLab
25+ years shipping production software, the last two on the front line of AI-augmented engineering. This course is the playbook I use myself, refined with engineering teams across three companies.
Newsletter
Practical Agentic Engineering.
Real-world lessons from adopting AI agents in software teams. From my own workflow. Every other week. No fluff.