AI-Native Software Development
Learning to build with AI as part of the team
AI is already part of software development.
The real question is whether teams are using it deliberately, safely, and in a way that actually improves their work.
This training helps software teams move from individual AI experimentation to shared, professional AI-native development practices.
We focus on how developers work with AI in real codebases: planning, coding, reviewing, testing, and iterating with humans firmly in control.
What problem this training solves
Many teams already use Copilot, ChatGPT or similar tools, but:
- results vary wildly between developers
- quality and consistency are hard to maintain
- AI use stays personal instead of becoming a team capability
This training builds a common baseline for AI-native development so teams can work faster without losing clarity, quality, or responsibility.
What teams learn
Participants learn how to:
- work with LLMs as part of everyday development workflows
- design prompts and context so AI output becomes reliable, not random
- use AI for code generation, debugging, reviews, and documentation — responsibly
- understand where AI helps and where human judgment is non-negotiable
- move towards agent-supported workflows without losing control
All learning happens in the team’s own development context, not in abstract examples.
How we work
We combine:
- short, focused theory
- hands-on exercises in real repositories
- shared reflection on what worked, what didn’t, and why
Who this is for
- Software engineers
- Tech leads and senior developers
- Teams preparing for more agentic or AI-assisted workflows
No hype. No silver bullets.
Just practical capability building for teams that want to take AI seriously.
Contents
Module I – Fundamentals
We cover the theory behind each topic, provide practical advice from experience, and apply it through
hands-on exercises which are done in participants own work context (code repository)
● Key concepts: LLM, MCP, RAG, Agents
Breakdown of all key concepts of AI native development.
● Prompt & Context engineering
We emphasize the importance of good prompt structure, using roles, and “chatting” with the LLM. It also introduces
the concept of using shared project rules to enforce conventions and ensure consistent outputs from LLMs
● Examples of various use cases for LLMs
Code generation, debugging, documentation, code reviews, and code base analysis.
● Tool and model selection
Familiarize participants with characterics of different tools and models
● Limitations of LLMs and the role of the developer, guiding rules
Responsible LLM use requires acknowledging its limitations, enforcing disciplined review of all generated code, and
validating the output with safeguards like peer reviews and automated testing.
Module II – Agentic Development
Goal of the Agentic Module is to learn to build development workflows that can autonomously plan,
execute, and iterate on tasks with minimal human intervention.
● Agentic Development Prerequisites
Set up Claude Code with proper project context, coding standards, clear templates, and quality checks
● Agentic Development Cycle
Learn workflows where AI agents autonomously analyze requirements, generate plans, execute code changes, validate
results, and self-correct based on feedback
● Tool Integration & Orchestration
Build agents that connect with your existing tech stack – IDEs, version control, CI/CD pipelines, and APIs
● Multi-Agent Collaboration
Design specialized agents with distinct roles – architecture, coding, testing, and deployment working in parallel
● Human-AI Governance & Oversight
Controls that define agent escalation points, establishing boundaries between autonomous and human decisions
When and where
Duration
2x 5h workshops and 2x 1h Ask me anything -sessions two weeks after the training
Location
Online or onsite, as agreed
Max 30 participants
Trainers
-
Ville Heikkinen
Technical Lead
-
Joonas Kykkänen
Technical Lead
Do you want to know more about our services?
Please send an email, or leave a message right here. We look forward hearing from you and get back to you asap.
marjut@splended.fi
