Course Objective & Scope
Using AI coding tools merely for line autocomplete or Blind copy-pasting ('vibe coding') quickly creates architectural drift and review bottlenecks in real production repositories. This 2-3 day advanced engineering lab teaches software teams disciplined AI-native engineering practices: spec-driven development, repository-level context rules and skills, bounded agentic execution, and automated verification loops.
Working hands-on inside a multi-layered codebase, participants practice orchestrating coding agents across the full software development lifecycle—from architectural planning and multi-file implementation to automated test generation, legacy refactoring, and high-signal code review.
Who Should Attend?
- Backend, frontend, and full-stack software engineers looking to master AI coding assistants and autonomous coding agents
- Tech leads, staff engineers, and engineering managers establishing AI-native SDLC standards and repository conventions
- Software architects leading legacy modernization, technical debt reduction, and test coverage initiatives
Key Takeaways
- Drive complex implementations using structured technical specifications (
spec.md, implementation plans) before generating code - Engineer repository context using instruction files (
AGENTS.md, rules, reusable skills) and Model Context Protocol (MCP) tools - Decompose multi-file features and complex migrations into atomic, verifiable subtasks for coding agents
- Apply AI-accelerated Test-Driven Development (TDD) to generate high-value unit, integration, and characterization tests
- Enforce rigorous verification, static analysis, security scanning, and review practices to prevent AI-induced technical debt
Curriculum
From Vibe Coding to Disciplined AI-Native Engineering
- Comparing inline autocomplete, chat-assisted coding, and autonomous CLI/IDE coding agents
- Diagnosing AI failure modes in production repos: architectural drift, bloated diffs, and subtle regressions
- The engineer's evolving role: context curator, specification author, and verification architect
Repository Context Engineering: Rules, Skills, and MCP
- Making repositories agent-ready with
AGENTS.md, architectural decision records, and scoped rule files - Packaging repeatable team patterns into reusable agent skills and custom slash workflows
- Connecting coding agents to live schemas, docs, and issue trackers via Model Context Protocol (
MCP)
Spec-Driven Development and Multi-File Execution
- The Research -> Specification -> Plan -> Execute loop for non-trivial features
- Locking down interfaces, data contracts, and edge cases before letting the agent write implementation code
- Closing the verification loop: wiring typecheckers, linters, and test runners directly into agent feedback
AI-Accelerated Testing and Legacy Refactoring
- Generating characterization tests to lock down behavior in untested legacy modules
- Practicing AI-assisted TDD: specifying failing tests first and guiding the agent to minimal passing code
- Executing safe multi-file refactorings, framework migrations, and dependency upgrades
Systematic Debugging, Security, and AI-Native Code Review
- Hypothesis-driven root cause analysis and log/trace investigation with coding agents
- Catching common AI security pitfalls: broken access control, injection risks, and hallucinated packages
- Maintaining small, reviewable PRs and combining automated AI review checks with rigorous human sign-off
Other Courses in This Category
-
A 1-2 day hands-on engineering workshop focused on building production RAG pipelines with document parsing, chunking, embedding, hybrid retrieval, reranking, and citation evaluation.
- Live Course
- 5 modules
-
A 1-2 day interactive tabletop simulation for executives and risk teams testing incident response, hallucination crises, data-leak scenarios, and EU AI Act compliance controls under pressure.
- Live Course
- 5 modules
-
A 0.5-1 day applied workshop on call transcript analytics, Voice of the Customer synthesis, intelligent intent routing, real-time agent-assist copilot design, and automated QA scoring.
- Live Course
- 4 modules
-
A 1-2 day sector-focused workshop applying AI and automation to demand forecasting, dynamic route and dispatch optimization, proactive exception alerts, and warehouse document processing.
- Live Course
- 5 modules
-
A 0.5-1 day hands-on workshop on designing metric trees, separating North Star, leading, and lagging indicators, refactoring cluttered executive dashboards, and establishing decision rituals.
- Live Course
- 5 modules
-
A 1-2 day immersive role-play and case simulation lab where managers practice high-stakes feedback, tough stakeholder negotiations, conflict resolution, and coaching conversations with live feedback.
- Live Course
- 5 modules
Get in touch
