Kurser til virksomheder og offentlige organisationer

This two-day course introduces participants to the fundamentals of agentic AI systems that can reason, plan, and act using large language models. Participants leave with a practical toolkit for working with agentic AI: not just understanding how these systems work, but also how to design and build simple agents, apply common patterns, and anticipate common challenges.

Day 1: Foundations of AI Agents
Day 1 introduces the core concepts behind agentic AI systems and quickly moves into practical application. You’ll learn how agents use reasoning, memory, and tools to solve tasks, and how design patterns structure their behavior. These concepts are applied in hands-on exercises using no-code tools.

By the end of the day, you will be able to:
- Understand the core components of agentic AI systems
- Apply basic design patterns to structure agent behavior
- Translate real-world tasks into step-by-step agent workflows
- Build and test a simple AI agent using no-code tools 

Day 2: Data, Limitations, and Building Reliable Agents
Day 2 focuses on foundations that make agents work in practice: well-structured data, robust design, and awareness of limitations. You’ll learn how the quality and structure of inputs, such as prompts, context, and documents, directly impact performance. The day also covers common failure modes, risks, and practical techniques to make agents more reliable, predictable, and safe. Through hands-on exercises, you will refine and improve your agents.

By the end of the day, you will be able to:
- Structure and prepare input data (prompts, context, documents) for agents
- Diagnose common failure modes in agent behavior
- Improve reliability through better prompting, memory use, and iteration
- Apply practical safeguards to reduce errors and unintended outputs
- Build more robust agents that perform consistently on real tasks

Dates: Thursday 8th & Friday 9th October

Workload: The total number of hours for this course is 16 hours including teaching and exercises.

Teacher: This course is taught by Assistant Professor Pepa Kostadinova Atanasova

Language: This course is taught in English

Price: 10.000 kr.

Kunstig intelligens fylder stadig mere i både samfundet og hverdagen. Teknologien bevæger sig hurtigt, mulighederne er enorme, og konsekvenserne – både positive og negative – er vidtrækkende. Dette kursus giver deltagerne et solidt, nuanceret og praktisk funderet overblik over kunstig intelligens, dens historie, dens påvirkning af mennesker og arbejdspladser samt de fremtidsperspektiver, vi allerede nu skal forholde os til.

Kursusforløbet består af tre sammenhængende seminarer, som kombinerer viden, refleksion, diskussion og konkrete analyseværktøjer.

Del 1: KUNSTIG INTELLIGENS OG SAMFUNDSVÆRDI:
Første seminar giver deltagerne en forståelse for, hvor kunstig intelligens kommer fra, og hvilke forestillingerne forskerne har gjort sig om teknologien de sidste firs år. Deltagerne får også et analytisk værktøj til at forstå, hvordan kunstig intelligens kan ses i forlængelse af tidligere teknologier, og hvordan kunstig intelligens kan skabe værdi i samfundet og i virksomheder, samt hvor man skal være mest opmærksom under udrulningen af kunstig intelligens. 

Del 2: KUNSTIG INTELLIGENS OG MENNESKER:
Alle teknologier påvirker, hvad det vil sige at være menneske. Og teknologier påvirker os forskelligt. Med udgangspunkt i første seminars afsluttende diskussion fokuserer vi på, hvordan kunstig intelligens påvirker os, når vi går på arbejde og i vores privatliv. Hvordan sikrer vi os, at vi ikke udliciterer ting, der er vigtige for medarbejdstilfredshed og kognitiv sundhed? Hvad og hvem skal vi bekymre os for? Og hvor er der nye muligheder? Vi diskuterer også, hvordan teknologi kan designes til at modvirke ulighed og deskilling. 

Del 3: KUNSTIG INTELLIGENS OG FREMTIDEN:
Kunstig intelligens er under hastig udvikling, og mens virksomheder og privatpersoner vænner sig til den nyeste teknologi, dukker andre teknologier op i horisonten. Hvor er kunstig intelligens på vej hen? Hvordan vil de påvirke arbejdsmarkedet? Og hvordan kan vi bedst forberede os - som samfund, virksomheder og privatpersoner? 

Underviser: Dette kursus bliver undervist af Anders Søgaard

Pris: 7.500 kr.

Obs: Hvis der er færre end 15 tilmeldte, forbeholder vi os retten til at aflyse kurset.

This course enables engineering teams to become productive with Python as a core platform for delivering data-driven and machine-learning solutions. By the end of the course, participants can confidently build, maintain, and iterate on end-to-end data and ML pipelines that are ready for real-world use. They also learn how to turn raw data into deployable insights by using Python’s data science and machine-learning ecosystem to ingest, clean, preprocess, and model data.

After participating in this course, your team will be able to:

• Train and optimize models with modern frameworks and automatic differentiation for faster experimentation
• Apply skills directly to real-world datasets, including participants’ own organizational data
• Achieve measurable impact through project-based work supported by curated example datasets

Prerequisites: If you are uncertain about complying with the prerequisites, please contact us for an evaluation.

• Basic programming in Python
• Linear algebra (1st year university level)
• Basic concepts of statistics (1st year university level)

Teacher: This course is taught by Associate Professor Oswin Krause

Price: TBA

Note: If there are less than 10 participants, we reserve the right to cancel the course.

This part builds on the foundations with advanced topics and hands-on implementation. You will deepen your understanding of evaluation, safety, and multi-agent systems, then build a working agent in one of two tracks.

Track A: Building Agents with Code
This course is designed for developers and engineers who want to build agents programmatically.

Learning outcomes:

• Understand agent evaluation methods, including benchmarks, tracing, and automated testing.
• Learn safety and alignment techniques: guardrails, human-in-the-loop, sandboxing, and output validation.
• Study multi-agent coordination patterns: delegation, negotiation, and shared memory.
• Build agents programmatically using LangChain, LangGraph, or CrewAI.
• Implement structured agent-tool communication using the Model Context Protocol (MCP).
• Deploy and monitor agents in production infrastructure.

Track B: Building Agents with No-Code Platforms
For professionals who want to build agents with drag-and-drop frameworks.

Learning outcomes:

• Understand agent evaluation methods: how to measure success, spot failures, and iterate.
• Learn safety principles for production agents: approval flows, fallback handling, and scope constraints.
• Study how multi-agent workflows are orchestrated in enterprise settings.
• Design agent workflows using no-code platforms (e.g., n8n, Microsoft Copilot Studio).
• Connect agents to enterprise tools such as CRM, ERP, and collaboration platforms.

På Datalogisk Institut kan vi hjælpe dig med at give dine medarbejdere et kompetenceløft som baseres på den nyeste forskning. Vores særligt tilrettelagte virksomhedskurser kan designes til virksomhedens specifikke behov indenfor emner såsom data management, machine learning, AI, softwareudvikling, chatbots & LLMs og meget andet. I kan finde mere inspiration i vores nuværende kursusudbud (scan QR kode). Vi kan afholde kurset hos jer, og vi tager gerne afsæt I jeres cases og data.

Vi har eksperter som kan hjælpe jer, hvis I har brug for at opkvalificere jeres medarbejdere til at opsætte en machine learning model, der kan håndtere jeres databehandling, udvikle/optimere en chatbot specifikt til jeres virksomheds behov, eller hvis jeres softwareudviklere har brug for en opdatering på det nyeste nye inden for computer science.