Crash Course: Agentic IDE for large, brownfield and legacy projects
Mentor - Vyacheslav Koldovkyy, Competence Manager at SoftServe
There is a common misconception that AI tools work better on greenfield projects than on existing, often large and complex, projects (brownfield). Indeed, a clean, well-structured codebase in a new project is ideal for AI analysis and code generation.
This does not mean that AI is ineffective for brownfield/legacy projects. It all depends on the approach.
- ❌ What doesn't work:
- Unsystematic use of AI
- "Vibecoding" without rules and context
- Random changes to the code without understanding the architecture
- ✅ What works:
- Documenting the codebase in an AI-friendly format
- Clearly defining the boundaries of the context
- Configuring rules and roles for agents
- Using Agent Skills for repetitive technical actions
- Selecting or developing your own MCPs
- Applying Spec Driven Development for safe changes
With the right approach, using AI in brownfield projects reaches a different level of productivity - often even higher than in greenfield.
In this course, you will learn how to effectively use the Agentic IDE for large, brownfield and legacy projects. During the practical tasks, you will have the opportunity to choose to perform tasks for a real open source project, or practice on your own project.
Language of the event: Ukrainian
For the demo, we will use Cursor, but in general, the approaches are universal and suitable for any Agentic IDE, in particular Claude Code / GitHub Copilot, etc.
Format:
Crash course: 3 online sessions
Dates:
March 24 and 26: start at 6:30 PM
March 28: start at 11:00 AM
Duration of each class: ~3 hours.
Platform: Zoom. You will receive a link to the broadcast before the course starts, and it will also be available on this landing page. In addition, all materials and lecture recordings will be available on the learning platform.
Preparation: Before the start of the course, participants will receive instructions for preparing the project and the working environment.
Homework: To consolidate the material and receive a certificate, participants must complete practical tasks that will be checked by a mentor.
Event language: Ukrainian
Presentation language: Ukrainian with English terms
Who will be interested?
- Developers of all levels and tech leads/architects, DevOps/SREs who want to stop “inventing prompts from scratch every time” and make interaction with agents predictable.
- QA/Automation – for describing validation rules, test generation and standardization of quality approaches.
- Mentors, team leads and Engineering Management specialists who build standards for working with AI tools in teams and companies.
Program
Day 1: Theory and practice of context management of large/brownfield/legacy projects
- What are the specifics of large, brownfield and legacy projects in terms of using AI
- The problem of context - why do agents fail on large repositories?
- What are the problems in working with self-written libraries, frameworks and legacy code?
- What is “Code Archaeology” and how to do it effectively
- How to measure the volume of a project in tokens
- How to properly document a project with AI
- How to make sure that AI does not break the project when making changes
- Features of migrating projects to new technologies
Day 2: Extending the IDE for a project: Rules, Agent Skills, MCP
- Rule architecture and composition, hierarchical rules for large projects
- How to describe architectural patterns of a legacy project so that AI does not offer "fashionable" but incompatible solutions
- Creating "Role-based" rules: how to make AI act as a Senior Developer of this specific project
- Protection from hallucinations: instructions "what NOT to do" (negative constraints) for old libraries
- The concept of Agent Skills: what is the difference between just a context and a skill
- Practice of creating skills: teach the agent to run local build scripts, deployment or specific tests
- How to describe skills so that the model selects them automatically at the right moment
- When Skills are not enough: connecting to external data, searching for MCP
- Developing MCP for your own tasks
Day 3: Safe Refactoring and Effective Spec Driven Development
- Why Pure TDD is Hard for Legacy, and How SDD Changes the Rules of the Game
- The SDD Cycle with AI: Plan → Spec → Implement → Verify
- Using MD Files as a Contract Between a Human and an AI
- Applying the BMAD Methodology to Orchestrate Specialized Agents
- Working with the Agentic IDE for Multi-File Refactoring
- The "Strangler Fig" Pattern with AI: Gradual Replacement of Old Code by New by Creating Parallel Structures Described in specifications
Mentor: Viacheslav Koldovskyi
— Competence Manager at SoftServe
— 20+ years in IT, certified Google Cloud Professional Cloud Architect, nVidia Generative AI LLMs
— Ph.D / Candidate of Economic Sciences, Associate Professor, Head of the Gen AI Center at IT STEP University
— Founder of YouTube and Telegram channels Programming Mentor
— Active speaker: speaking at iForum, DOU Day, etc.
— Leader of the AI community at DOU, author of publications
— Has implemented projects with Gen AI and AI-generated code, which successfully working in production
— Advises companies on transforming SDLC processes using AI
Event price
Access to 3 online broadcasts (~10 hours)
Presentations and materials
Access to the course recording on the learning platform (1 year)
Access to the chat in the TG, where the speaker will answer questions (1 month)
Participant certificate (subject to completing homework)
Free month in Fwdays Club for those who receive a certificate
Course participants receive a 10% discount on participation in Fwdays conferences
Affordable payment in installments from Monobank and purchase in installments from PrivatbankIf after the first lesson you realize that your expectations do not match, we will refund the money in full.