Videocourse: Agentic Engineering for new projects
Mentor - Vyacheslav Koldovskyy, Competence Manager at SoftServe
Why do some AI-first projects take off while others stay stuck at the demo stage?
The secret lies not only in the tools, but in how the engineering around them is built. Claude Code and OpenClaw are vivid examples of how a greenfield approach lets you get the most out of agents: lay down the right architecture, think through the rules of interaction, connect the tools you need and move towards production at an entirely different speed.
Mentor: Vyacheslav Koldovskyy, Founder Dev AI Consulting.
In this video course you will examine the approaches behind such AI-first solutions and learn to apply them in practice to build your own production-ready products.
Important: how do you work with AI agents properly?
- chaotic use of agents in «vibe-coding mode», a quick start with no considered architecture or optimal stack
- generating code without specifications or acceptance criteria
- agents working in parallel without coordination
- faster delivery paid for with technical debt
- no reliable guardrails and no review process
- a deliberate choice of technologies, frameworks and architecture for AI-first delivery
- building a system of rules, roles and skills for agents from the very start
- a balanced approach to connecting MCP for external tools and data sources
- Spec Driven Development as the basis for controlled implementation
- multi-agent orchestration for parallel work on the system
- an uninterrupted path from idea and specification to production
Done right, AI in greenfield projects gives you not just «faster code generation», but a fundamentally different pace of product creation: the team focuses on features, business logic and decisions, while agents take on a significant part of the technical implementation.
You will build a full-stack web project from scratch step by step, using the best modern Agentic Engineering practices, and embed an OpenClaw / NanoClaw-based agent into the solution. During the hands-on assignments you will turn your own idea or startup into a finished product. Cursor is used for the demos, but the approaches are universal — they work with any Agentic IDE, including Claude Code and GitHub Copilot.
After the course you will:
- be able to start a new AI-first project and deploy it to production on the very first day
- grow the product through Spec Driven Development with rules, skills, MCP and multi-agent orchestration
- keep quality up and avoid technical debt through guardrails, code review and fast feedback loops
- embed an OpenClaw / NanoClaw-based agent into your product and move individual features into the Agent Skills format
- speed up delivery to production without losing quality or control
- Developers of all levels who want to launch new products with AI through an engineering approach rather than chaotically
- Tech Leads, Solution Architects and Founding Engineers who design new systems for agents to work in effectively
- Startup founders and product builders who want to shorten the path from idea to working product and production deployment
- DevOps / Platform Engineers who want to build a fast delivery pipeline for agent-driven development
- Engineering Managers and team leads who want to standardise the use of AI tools on new projects
- Product managers and business analysts who want to implement their own ideas in production themselves
Three modules: from the first deploy to an agent in production
module-1 Shipping your first working product to production 1/3 ›
- – The new SDLC: why generation is no longer the bottleneck, but verification is
- – Agent = Model + Harness: what agentic engineering is made of — models and benchmarks, tools, rules, Agent Skills, MCP, sub-agents
- – Greenfield specifics: an AI-friendly stack, languages and frameworks, agent-friendly architectures
- – How Claude Code and OpenClaw are built — and what to take from them
- – Comprehension debt and intent debt: why «everything in the repository» and SDD are critical
- – Practice: from PRD to a deployed Next.js app — the naive vibe approach vs an engineering setup (React best-practices skill + Context7 + Vercel MCP)
module-2 Launching a project factory 2/3 ›
- – Project Factory: a repeatable cycle of «requirements → specifications → tests → implementation → verification → proofs»
- – Loop Engineering and its parts: automations, worktrees, skills, connectors, sub-agents, memory
- – Spec-Driven Development (OpenSpec) and capability slicing instead of «build the whole MVP at once»
- – Running agents in parallel in isolated worktrees
- – Guardrails and review gates: maker ≠ checker, evals, tracing and automated QA
module-3 Embedding an agent based on OpenClaw / NanoClaw / Hermes 3/3 ›
- – Skill-based programming in action: one of the features built on Day 2 with ordinary deterministic code is turned into an agent and integrated with the web interface — producing a hybrid «web + agent runtime» architecture
- – OpenClaw / NanoClaw / Hermes: a short overview of the approaches and when to choose which
- – Skill-based programming as a paradigm: when business logic is better off living as Agent Skills than as code
- – From a deterministic feature to an Agent Skill: rewriting a piece of Day 2 code in skill format
- – Integrating the agent with the web interface; optionally an additional channel through a messenger
- – Security, models, cost and limits: what to watch for when you put an agent into production
Vyacheslav Koldovskyy
- – 20+ years in IT, certified Google Cloud Professional Cloud Architect, nVidia Generative AI LLMs
- – Ph.D in Economics, Associate Professor, Head of the Gen AI Centre at IT STEP University
- – Founder of the Programming Mentor YouTube and Telegram channels
- – Active speaker: talks at iForum, DOU Day and others
- – Leader of the AI community on DOU, author of numerous publications
- – Has delivered Gen AI and AI-generated-code projects that run successfully in production
- – Advises companies on transforming their SDLC processes with AI
Event price
Access to recordings of 3 online course sessions
Presentations
Practical tasks for independent study
A certificate (provided you pass the test)
10% discount on participation in Fwdays conferences
Available payment in installments from Monobank and purchase in installments from PrivatbankAfter purchase, we will send you access to the course within one business day