Development with AI tools
Vibe coding and AI-assisted development: defining tasks, working with repository context, generating and reviewing code.
Online / In person / Hybrid
We teach IT teams to apply AI tools in development, build AI agents and create LLM-based applications. Practice centres on engineering tasks, architecture and company requirements.
View the programmeVibe coding, AI agent development and AI application creation can be studied separately or in one programme.
We work with repositories, APIs, data and architecture close to your team's tasks.
We cover testing, review, quality evaluation and operational requirements.
Content is agreed around languages, tools, model access and security requirements.
We build the programme around the processes your team works with every day.
Vibe coding and AI-assisted development: defining tasks, working with repository context, generating and reviewing code.
Designing agent workflows, API integration, state management and action control.
Document preparation, retrieval, context assembly and answer validation using internal data.
Building interfaces and backends, connecting models and developing user scenarios.
Quality evaluation, test sets, access control, untrusted data handling and secret protection.
Deployment, tracing, error monitoring, and model latency and cost control.
We adapt the programme to participants' roles and departmental needs.
Find a programmeFrontend, backend and full-stack engineers introducing AI tools into development.
Specialists defining AI solution architecture, standards and integrations.
Engineers working with data, retrieval, models and quality evaluation.
Specialists in application validation, deployment and support.
The programme aims to embed AI in development and operations processes and establish practical workflows within the team.
The team establishes rules for AI tool use, code validation and change management.
Participants design agents with tools, constraints and verifiable behaviour.
The team progresses from a user scenario to an application with quality evaluation and integrations.
Solutions receive operational, monitoring, security and cost control requirements.
Choose one track or combine them into a progressive programme. We agree on content and depth based on the team's experience and technology stack.
AI-assisted development under engineering oversight: from defining a task to a verified repository change.
Requirements, decomposition, codebase context and coding assistant rules.
Interface and API prototyping, working with existing code, error analysis and iterative refinement.
Diff review, tests, reviews, dependencies, secrets and preparing changes for CI.
A verified change or prototype in a training repository, tests and AI tool usage rules.
Designing agents that use tools and perform bounded actions within workflows.
Task, state, tools, structured output and boundaries for autonomous actions.
RAG, document search, API calls, access and external system error handling.
Human confirmation of significant actions, tracing, test scenarios and protection against instructions in untrusted data.
An agent prototype with tools, access limits, an execution log and test scenarios.
Building applications with LLM features and preparing them to operate on company infrastructure.
User scenario, interface, backend, models and corporate data.
Test example sets and evaluations of quality, latency and request costs; error and failure handling.
Authentication, configuration, CI/CD, observability, limits and application update procedures.
A working training prototype of an AI application with an interface, API, quality evaluation and deployment plan.
Before training, we study the team's processes and tasks, employees' existing experience and company requirements. These inform the content, tools and practical work.

We establish which processes, roles and outcomes the team wants to improve with AI.
We gather typical tasks, existing tools and employees' current AI experience.
We select content, tools and practical work for specific roles and participant skill levels.
Participants work with repositories, connect APIs, develop agents and AI applications, evaluate quality and prepare deployment plans.
We document workflows, templates and the next processes to automate.
Participants learn the tools and apply them to the tasks they work on every day.
We define specific exercises after learning about the company's needs.
Prototypes are developed in the agreed environment. Production system access and release follow normal review and acceptance procedures.
Make a repository change with an AI assistant and review the diff
Develop an agent that calls an external API
Build search and answers over a document set
Create an AI application with an interface and backend
Prepare test examples and measure quality
Configure tracing, limits and a deployment plan
Format and duration depend on the objectives, participants and depth of practical work.
For exploring AI capabilities or addressing a specific work task.
For mastering a particular category of tasks and reinforcing skills through practice.
For progressively developing skills and embedding new ways of working into team processes.
Three standalone tracks: vibe coding, AI agents and AI applications. We adapt depth and stack to participant roles and existing infrastructure.
Online / In person / Hybrid
Depending on the programme, your company retains:
Programmes vary in audience, depth and objectives, from leadership training to technical programmes for developers.
Fashion retailSenior management · online
Launch cascading AI implementation: from the leadership team's strategy to programmes for every department.
Senior leaders identified implementation opportunities, and the programme is then cascaded to departments.
Diversified holding companySenior management and legal teams · in person
Learn to use AI safely in legal processes as the first stage of a broader implementation programme.
The legal team uses AI safely and has built its first prototypes for its processes.
IT integratorLeaders and 5 technical teams · online
Take teams from leadership strategy to hands-on AI agent development.
Teams created a list of AI initiatives and built AI agents for them.
Software developmentDevelopers and team leads · online
Understand the inner workings of AI coding assistants as well as using ready-made tools.
Developers built their own coding agent for their technology stack.
International trade and logisticsSales and marketing · online
Give the sales department practical tools for its tasks, beyond general knowledge about AI.
The sales department launched its own AI agents and no-code automations.
TravelTechEntire company · 3 cohorts · online
Train the whole company at once despite substantial differences in employees' existing skills.
Every employee learns in a cohort suited to their level, from basic literacy to multi-agent systems.
We will review your employees' needs and experience, then propose a format, programme and training cost.
Find SkaiLab on LinkedInVibe coding covers development with AI tools and code validation. The AI agents track covers tools, state, integrations and action control. The AI applications track includes interfaces, backends, quality evaluation and operations. Tracks can be taken separately or sequentially.
Agent and application development require programming and API experience. The AI-assisted development track accounts for participants' existing engineering experience. Entry requirements and group allocation are agreed before training.
Yes, if company rules allow it. We clarify access, approved tools and code transfer restrictions. Where necessary, we use a separate repository and training data.
Training produces a prototype or a verified change in the agreed environment. Production release requires the company's standard checks, review, security assessment and acceptance.
Before launch, we study participants' tasks, skills, workflows and tools. Content and practice are adapted to the specific team.
No. Describe the tasks and team composition. After an introductory meeting, we will propose content, format and duration.
Yes. Where necessary, we conduct an initial assessment and divide participants into cohorts by role or skill level.
The stack is determined by company tasks and constraints. If approved services and infrastructure already exist, we adapt the programme to them.
Yes, within agreed access and information processing rules. Where needed, we use anonymised documents, training data and a separate test environment.
We use practical assignments, final projects and feedback. For prototypes, we define test examples and quality criteria. Evaluation scope is agreed before training.
Online, in person or hybrid. Options include a standalone workshop, a session series or a programme for several cohorts.
Cost depends on participant count, duration, customisation depth and format. After an introductory meeting, we prepare the programme and cost estimate.
Tell us about your team's needs. We will suggest a programme, format and training duration.
Get the programme