Johan Kruse
Aarhus, Central Denmark Region, Denmark
1K followers
500+ connections
About
Being ambitious is key - let's go together!
Team lead key points:
-…
Activity
1K followers
-
Johan Kruse shared thisJeg har snakket med tre den inden for den seneste uge, som var lidt flove over den. Du kender den godt. Den, du har bygget sammen med Claude. Ingen andre kan tilgå den, men den kan gøre din hverdag markant bedre. Specielt hvis den får fat i de rigtige data. Jeg er kæmpe fan af de løsninger, for - du ved, at problemet kan løses, - du ved, hvilke data der ligger under, og hvor der er mangler, - du ved, hvordan den reelt bruges (ikke helt som den er designet!). Nu skal løsningen rettes til, sættes i drift og måske tåle mødet med en kunde eller kollega kl. 03 om natten. Du er ikke alene. Og vi hjælper en hel del som dig videre, når det føles som lidt af en forhindringsbane..
-
Johan Kruse shared thisNæsten månedligt vender en gammel kollega “hjem” i øjeblikket. Det er tegn på en kultur man gerne vil være en del af 😉 Og vi vil selvfølgelig hellere møde dig, end at have en AI-genereret ansøgning. Ræk ud og lad os bygge noget fedt!Johan Kruse shared thisVi søger Principal Data Scientists. Og Senior Data Scientists. Og Data Scientists, AI Engineers, Forward Deployed Engineers, ML Engineers. Taksonomien for titler skifter lige så hurtigt som modellerne. Jeg regner med at du også har opgivet at følge med, og at vi i stedet skal snakke om, hvad du har bygget. 🤖 Det, jeg leder efter, er svært at få ned i en jobannonce. Så jeg prøver sådan her: Du bliver irriteret, når noget tager 120 timer, og du kan se, at de 60 er spild. Og når det åbenlyse svar ikke virker, bliver du nysgerrig i stedet for træt. Bagefter kan du fortælle en kunden, hvorfor det betyder noget. "Vi gjorde det 50% hurtigere at svare på udbud" er en bedre sætning end noget som helst med RAG i. Vi har flere projekter, end vi kan nå. Aarhus og Lyngby. Kender du hende eller ham? Tag dem. Skriv til Steffen, Martin (Aarhus) eller Rosa (Lyngby). Eller til mig.
-
Johan Kruse shared thissådan - Anthropic har valgt os som en af de første europæiske Select Partners. det betyder tættere adgang til "Claude behind the scenes", men også en anerkendelse af hvad der allerede leveres. vi kører upskilling hele ugen på tværs af landegrænser. Det kan godt gøre ondt på gamle vaner og komfortzoner. Måske burde det også gøre lidt ondt hos jer?
-
Johan Kruse shared this500 mennesker samlet til AI & Data Day for to uger siden. At tale med sine data dukkede op igen og igen, for det virker faktisk nu. PensionDanmark, VELUX og COWI og andre var på scenen og viste gode eksempler. Men hvad betyder det for klassisk rapportering? Hvornår er et dashboard den forkerte løsning? Hvad kræver det af vores data at implementere? Det går vi i dybden med til webinaret 3. juni. Der er ca. 400 tilmeldte allerede. Link i kommentar 👇
-
Johan Kruse shared this"Cloud only" - skrev vi i vores strategi i 2020. Men... Dengang var det det rigtige valg. Cloud har givet os hastighed, skalerbarhed og fleksibilitet, vi ikke ville have haft ellers. Ingen fortrydelse der. Men AI udfordrer nogle af de antagelser, vi byggede det på. De seneste måneder har gjort AI-infrastruktur mindre forudsigelig. Usage limits, dyrere workloads og kvalitetsproblemer fylder. Selv nogle nye topmodeller får et lidt lunkent “golf clap”, når de lanceres. Og vi begynder at se tegn på bevægelse den anden vej. Mac Minis hos vores kunder. Virksomheder i finans, forsvar, telco og retail, der aktivt vælger delvist on-prem. Ikke fordi de alle er tvunget til det, men fordi det giver bedre mening. Og Apple peger i samme retning med de nylige udskiftninger. Apple Intelligence er jo ikke noget man bruger frivilligt – men det er tydeligt hvor det er på vej hen. Det trækker paralleller til Apple II der flyttede compute fra mainframes og ud til brugeren. Vi storforbruger cloud AI mens vi stadig kan, men økonomien hænger ikke sammen for virksomhederne bag. Noget ændrer sig snart. I dag bygger vi ikke til on-prem. Det giver ikke mening. Men det er en snak, der går igen her – på tværs af lande, teams og kunder. “Cloud only” – måske skal vi sige farvel til den vane snart?
-
Johan Kruse shared thisField notes from 10 CxO conversations in 10 days across Germany, Denmark and Greenland. AI is high on the agenda, but value is very uneven. 7 out of 10 have “AI” as a top priority in some form. 5 out of 10 see no meaningful time savings yet. 5 out of 10 urgently lack the right people. 2 out of 10 have already lost significant money on AI plans that did not deliver. 3–5 out of 10 are revisiting their business strategy because important new opportunities have opened up in the last six months. The framing varies quite a bit depending on who you talk to. CEOs, CIOs and CCOs often talk about growth: “What can we do to win the race?” Many others still talk mainly about efficiency: “How do we run faster?” Not a huge surprise, but it matters. One framing pulls AI into strategy, business models and the value chain. The other often keeps it closer to smaller efficiency gains. Two sentences kept coming up: “If we just had the data..” “If we just had a more streamlined development process..” Both are fair. Both matter. But they can also become very convenient reasons to stand still. It is a bit like waiting to become rich before you start saving. The logic is backwards! And maybe this is a much needed shift for the people trying to fund the foundational work. For years data platforms, architecture and development processes have often been hard to sell internally because they sound like cost, complexity and plumbing. Now there may finally be a more compelling reason to fix them. AI needs better data, of course. But agentic development also changes the conversation around how fast we can build, improve and operate data platforms. That makes the opportunities easier to see, and the business cases easier to discuss. Another interesting pattern: buy vs build is not really binary. 3 were mostly building. 3 were mostly buying. 4 were doing both at the same time. Which might be where many companies end up. So maybe it is less about: “Should we buy or build?” And more about: “Where do we need to own the capability to create an advantage?” It was an interesting couple of weeks. Some thoughts were confirmed, others changed. Now I’m looking forward to getting home to my family — and hopefully not hearing the word “agentic” again until Monday. PS. One CEO in retail has chosen the “I’ll let SaaS and consultants make me AI-enabled while I focus on what I’m actually good at” strategy. Respect.
-
Johan Kruse shared thisLandet i Nuuk 🇬🇱 De næste dage møder jeg en række grønlandske virksomheder, med det fokus at få AI til at arbejde for dem. Ikke som et buzzword, men med udgangspunkt i det der faktisk driller. Flaskehalse i driften, strategiske muligheder der endnu ikke er i spil og tidskrævende processer. Jeg har stadig et par huller i kalenderen i denne uge, hvis I vil ønsker gratis inspiration. PS. Landede i t-shirt til vandret snevejr. Begynderfejl. Passende metafor for mange af de AI-strategier vi ser i øjeblikket. Det skal nok klare op 🙌
-
Johan Kruse shared thisThe old currency was hours. The new currency is outcome. On our way to Frankfurt to discuss the EU market approach with valantic colleagues from all around. Looking forward to doing things a bit differently! I brought a ton of implemented AI cases and tools and hope to prove that tech is not stopping the show. It’s all about deciding what to solve and actually following through. Old news, new opportunities.
-
Johan Kruse shared thisMødelokale booket. Én deltager. Nja, to faktisk, hvis man tæller Claude med. Jeg undrer mig over at det stadig er akavet at tale med mine agenter, men ikke mine kollegaer. Samme plads, samme skærm. Vi rykker lige nu rundt på kontoret for at samarbejde bedre. Åbent, lyst, plads til alle — mennesker. Jeg tager en walk and talk med min PA og Whispr. Den har ikke afvist en invitation. Endnu..
-
Johan Kruse liked thisJohan Kruse liked this"Vi mister evnen til at tænke, når AI overtager en opgave." Eller gør vi? Mit feed oversvømmes i øjeblikket af frygten for AI. Vi mister vi ingenting. Vi flytter, hvor vi bruger vores evner. Og det har vi gjort før; fra omstillingsbordet med telefondamer, til GPS-en der omdirigerer dig i trafikken. Hvert lag af manuelt arbejde, vi fjerner, flytter vores opmærksomhed op ad kæden, til der hvor vi stadig kan afgøre, om noget er rigtigt eller forkert. I en økonomiafdeling, brugte en controller flere timer om ugen på at trække rapporter ud af Ibistic. Bekymringen var, at indsigten i, hvad der faktisk stod i rapporterne, gik tabt, hvis vi automatiserede det hele. I stedet fik han mere indsigt: kunne filtrere, sortere på flere dimensioner og lave dybdegående analyser. Arbejdet flyttede sig, fra at trække data til at prioritere de mest kritiske sager og rådgive resten af virksomheden bedre. Tilliden er det, der afgør, om vi tør flytte os opad. Sjældent handler den om, hvorvidt resultatet er korrekt. Den handler om, hvorvidt vi bagefter kan gå ét led tilbage og spørge: hvorfor gør vi det her, og hvilken værdi skaber det? Det samme spørgsmål gælder, når det er en AI-model, der leverer en anbefaling, og ikke en controller, der trækker en rapport. Det, der oftest holder virksomheder tilbage fra at komme i gang, er frygten for forandring, særligt fra medarbejdere, der frygter at miste deres job. Men det, de færreste sætter ord på, er hvad jobbet egentlig indeholder: retten til at være den, der kan afgøre, om noget er rigtigt eller forkert. Den ret forsvinder ikke, når vi flytter os opad i beslutningslaget. Den flytter sig ligesom hos controlleren. I stort set samtlige virksomheder, jeg har arbejdet med, har det vist sig, at de har leveret mere med samme kapacitet eller væækstet, og medarbejdernes viden og ekspertise er blevet flyttet til et højere abstraktionslag, hvor de kan skabe mere værdi. Så hvad er vigtigst? At bevare måden, vi arbejder på, eller skabe værdi med den tid, vi får tilbage? Vi mister ikke noget, vi får flere muligheder. (billedet er fra min rejse i Vietnam, hvor jeg skulle have tillid til at andre ville hjælpe mig) #AI #Transformation
-
Johan Kruse liked thisJohan Kruse liked thisHvordan skalerer man Data & Analytics i en retailforretning med 25 forskellige brands? Det er ét af de perspektiver, vi tager op på Data & AI in Retail, når vi den 10. november samler retailbranchen i København til en halv dag med konkrete cases, faglige indlæg og erfaringsudveksling. En af dagens cases kommer fra DK Company A/S, hvor Simon Mundt Linulf, Director of Data & Analytics, deler sine erfaringer. For når efterspørgslen på data, indsigter og nye løsninger vokser, følger der også en række spørgsmål med: Hvad skal prioriteres? Hvad giver mening at løse fælles på tværs af 25 brands? Og hvad skal forretningen selv eje? Simon deler erfaringer fra DK Company med at positionere og skalere Data & Analytics i organisationen - og hvorfor de spørgsmål kun bliver mere relevante i takt med, at AI skaber nye muligheder. 📅 10. november 2026 📍 Charlottehaven, København 🎟️ Gratis deltagelse Se hele programmet og tilmeld dig: https://lnkd.in/e5UzGtET Tag endelig en kollega eller to med under armen, hvis I er flere, der arbejder med data, AI eller digital transformation i jeres forretning.
-
Johan Kruse liked thisJohan Kruse liked thisMit område hos Inspari har i mange år heddet Cloud & Infrastructure. I dag hedder det Platform Engineering. Det kan lyde som et navneskifte, men for mig afspejler det faktisk en ret stor udvikling i den opgave, vi løser. Platforme er ikke længere bare noget, der skal være stabile, sikre og tilgængelige. De er blevet en langt vigtigere del af virksomheders evne til at udvikle sig. - Til at skabe overblik. - Til at arbejde mere skalerbart. - Og til at tage ny teknologi i brug på en måde, der også fungerer i praksis. For mig er det netop essensen af Platform Engineering: Building the foundation that enables everything else. Det handler om at bygge det fundament, som gør det lettere for organisationer at udvikle, skalere og skabe værdi - også når nye krav og nye teknologier melder sig. Og lige nu er AI selvfølgelig en stor del af den udvikling. Det er også noget af det, jeg hæfter mig ved i vores AI at Scale-rapport: Når AI skal bevæge sig fra ambition og eksperimenter til reel anvendelse i organisationen, bliver fundamentet omkring teknologien mindst lige så interessant som teknologien selv. 👉 Læs AI at Scale her: https://lnkd.in/ezeCrJ_g Det betyder også, at mange af de diskussioner, vi i årevis har arbejdet med inden for cloud og infrastructure, får en ny relevans. Det er den udvikling, jeg gerne vil dele nogle perspektiver på her den kommende tid. Ikke AI set fra modellernes eller use casenes perspektiv. Men set fra den platform, der skal få det hele til at fungere i praksis.
-
Johan Kruse liked thisJohan Kruse liked thisKan du tale med dine data? Interessen for Conversational BI og Agentic Analytics vokser hurtigt. Allerede nu har mere end 65 deltagere fra både enterprise-virksomheder og offentlige organisationer meldt sig til vores briefing i Lystrup den 8. oktober. På dagen dykker vi ned i, hvordan Conversational BI fungerer i praksis og vi deler perspektiver på governance, arkitektur og skalering. Vi kommer også omkring vigtigheden af adoption og enablement ifm brugen af nye AI-værktøjer. Du møder derudover Mia Quist fra furnipart, som deler deres erfaringer med en Conversational BI-løsning bygget på Microsoft Fabric og Databricks, der giver medarbejdere adgang til salgsindsigter direkte via Microsoft Teams. Er du ansvarlig for Data, AI eller Analytics, er dette en oplagt mulighed for at få inspiration til, hvordan Conversational BI kan skabe reel forretningsværdi. Læs mere og tilmeld dig her: https://lnkd.in/eJxx6iFk
-
Johan Kruse liked thisJohan Kruse liked this🔔Career update🔔 Change is inevitable and some changes will have a greater impact on your life than others… …and career changes probably fall into that category! I want to share that I’ve resigned from my position at Inspari A/S to pursue something new! 👀 People have an extraordinary ability to leave an impression on you and for this I want to be thankful: 🔹Rosa, thank you for our great conversations - professionally and privately. While many talk about the technical side of AI, we usually discussed the human side of the scale! 🔹Philip, thank you for bringing me in and shared my passion for business-oriented AI & Automation transformation! 🔹Philipp, thank you for being one of the most positive human beings I’ve met and for a kickass time together on projects! 🔹Fabian, thank you for welcoming me on a remarkable European project - one I would not have been without! 🔹Rasmus and Emil, thank you for the sparring and guidance on deliverables - pushing the boundaries while remaining pragmatic and realistic! 🔹…and many more! As I wait for the new chapter to begin (more on this in November), I will enjoy the upcoming and final days of wedding preparations as well as a wonderful honeymoon in Asia 😍🙏
-
Johan Kruse liked thisJohan Kruse liked thisDen 29. september deltager jeg som oplægsholder på Computerworlds konference Fra teknolog til strateg: IT-ledelse i en AI- og forretningsdrevet virkelighed. Et af de spørgsmål, der fylder meget i dialogen med vores kunder lige nu, er: Hvordan omsætter vi AI-ambitioner til målbar forretningsværdi? Mange virksomheder har allerede idéerne og de første AI-piloter. Udfordringen er ofte at få AI forankret i processer, arbejdsgange og beslutninger, så det bliver en del af den måde, virksomheden faktisk arbejder på. Det er også et centralt tema i vores AI at Scale-analyse. For mig handler AI i stigende grad ikke kun om teknologien, men om prioritering, adoption og evnen til at skalere det, der skaber værdi. Det glæder jeg mig til at diskutere på konferencen. Hvis emnet også interesserer dig, kan du stadig nå at tilmelde dig 👇 https://lnkd.in/eYEQAFqW Samtidig er AI adoption er også et centralt tema i vores AI at Scale-rapport, som du kan læse her: https://lnkd.in/ejCUGeYR. #AI #Inspari #ITledelse #DigitalTransformation #BusinessValue
Experience
Education
Licenses & Certifications
Publications
-
Positronium formation from porous silica in backscattering and transmission geometries
European Physical Journal D: Atomic, Molecular, Optical and Plasma Physics
The Aarhus positron beam line is utilised to measure the positronium formation fraction from porous silica thin films created by the Glancing Angle Deposition technique. The highest formation fraction found from these studies in a backscattering geometry is (57.7 ± 1.0)% in good agreement with other measurements. In transmission mode, the maximum positronium output is found to be (12.5 ± 0.5)%. These are the first measurements of positronium formation in transmission of a porous silica thin…
The Aarhus positron beam line is utilised to measure the positronium formation fraction from porous silica thin films created by the Glancing Angle Deposition technique. The highest formation fraction found from these studies in a backscattering geometry is (57.7 ± 1.0)% in good agreement with other measurements. In transmission mode, the maximum positronium output is found to be (12.5 ± 0.5)%. These are the first measurements of positronium formation in transmission of a porous silica thin film, a starting point for future attempts to optimise the positronium formation in transmission.
Other authorsSee publication
Languages
-
English
Full professional proficiency
-
Danish
Native or bilingual proficiency
-
German
Limited working proficiency
-
Norwegian
Professional working proficiency
Other similar profiles
Explore more posts
-
Fredrik Ring
A few weeks ago I had a chat with a colleague, Eliza Nordén, and Tom Nilsson, co-CEO of Icebug, about Icebug's data-driven inventory management. It's become a core part of how they run their business. An interesting challenge came up. Between seasons - spring/summer and fall/winter - the system loses accuracy. Old season forecasts are in their tail end, new ones are just ramping up. And real-world demand doesn't switch cleanly between seasons the way a model assumes. My instinct was to solve this with better modeling. Smooth the transition, blend the forecasts and add complexity. But talking it through, we landed somewhere else entirely. In that gray zone between seasons, they simply don't need the system to the same degree. The pieces for the new season are already set. By the time they have to start shuffling stock, the new forecasts have full accuracy again. So instead of fixing the gap, we left it. At least for now. Evidently, not every dip in model performance is a problem to solve. This threw me off. Data systems exist to support decisions, not the other way around. It's worth pausing before optimizing and asking "what benefit will the effort actually bring?". Technical model performance and business value aren't always the same thing. This is part of my core message when doing lectures on the topic, but it's still a lesson I keep learning.
2 Comments -
Karin Szilágyi
The Card Visual is dead - long live the Card Visual! If you’ve recently installed the Power BI November update, you might notice something a little off in the Visual Pane: the Card visual might be missing! The “new” Card visual is now GA, fully replacing the “old” Card visual. Several users in the Community Forum have reported that all their card visuals disappeared after updating (it happened to me too). If you’re seeing the same issue, you can bring the (new) Card visual back by selecting “Restore default visuals” in the Visualizations Pane. Your previously created cards might also look a bit wonky after the update due to the new default style settings. Here’s an article that explains the changes and how to fix some of the formatting issues: https://lnkd.in/d8tsSqY2 Power BI November Update Summary: https://lnkd.in/dBtFUWQw Microsoft Learn - Create a Card Visual in Power BI: https://lnkd.in/dWPHzEHy How do you feel about this Update?
34 Comments -
Laszlo Molnar
📊 Need your Power BI data in Excel — without losing formatting or spending ages cleaning it up? Exporting from Power BI to Excel sounds simple… but if you just hit “Export data”, you often get raw tables that are hard to use. The good news? There’s a smarter way that keeps your rows, columns, and structure intact — especially when you want to share reports or hand over data to someone who lives in Excel. In our latest guide, we break down how to export from Power BI to Excel while preserving your valuable structure — not just a raw dump of numbers. 🔧 🔍 What you’ll learn: • Why the simple “Export Data” button often isn’t enough • The difference between exporting underlying vs summarized data • How Power BI’s export behavior affects formatting • Practical steps to get useful Excel output from your dashboards 🔗 Read the full step-by-step guide: 👉 https://lnkd.in/d6Fy2DWB Whether you’re finalising reports for clients, handing off to leadership, or just want clean, usable data from your dashboards — this walkthrough makes exporting predictable and practical. Stop wrestling with messy Excel exports… and start getting the data you actually need. 🧠🔁 #PowerBI #Excel #DataExport #ReportingTips #Connectorly #BusinessIntelligence #Analytics
-
Anna Bergevin
I cannot figure out why more people aren’t talking about Power BI moving towards BI as code. This is a big deal! Is it still janky? Yeah, okay, it’s still janky. Does it seem like a hack - yeah, it does. Is it true BI as code - yeah…this is like more like an MVP that’s just barely viable. But can you change your dashboard via code workflows? Yes. Could you use agents and natural language prompts to modify your dashboards via these files —-YESSSSSS! Now I just want Microsoft to release 1 bajillion times faster to reduce the janky factor. Like I don’t want to even have Power BI desktop installed at all…bay steps.
17 Comments -
Reid Havens 🧙🏻♂️
📢 Good news everyone! ⭐ As of March 2026 PBIR is now the default format in Power BI Desktop. Power BI source control has had an amazing but iterative journey over the years. Mathias Thierbach helped kick it off with PBI Tools and TMDL, and now were getting a better default structured file format for even the report layer ⚠️ Previously, you'd export a PBIX, maybe decompose it with third-party tools (like PBI Tools), pray your merge conflicts were manageable, and pretend that counted as CI/CD. We all knew it was duct tape. Functional duct tape, sure, but duct tape. ⌚ The March 2026 release flips PBIR on as the default in Desktop. Meaning your report definition lives in a format that plays nice with Git, with deployment pipelines, with scripts, and with LLMs that can more easily read and edit report structures programmatically. 📝 Now, if you're a solo report builder who just publishes to the service and moves on, this probably won’t change your development landscape much. But if you're on a team, if you've ever lost work to a PBIX overwrite, or if you've wanted to code-review a report change…this is the foundation for all of that. 💡 This doesn’t mean PBIX is going away, PBIR will exist as a format BOTH in PBIX’s and PBIPs going forward. Either way this is the right evolution going forward. #PowerBI #Fabric #PBIP #TMDL #ProDeveloper
7 Comments -
Deepesh Vashistha
I’ve worked with some really smart Power BI / Fabric developers. And I’ve also seen why two people with the same tools… end up with very different outcomes. The difference is rarely DAX or visuals. A mediocre dev focuses on: “Tell me what to build.” A good dev quietly worries about: • where this data is coming from • what will break when volume doubles • why this number will be questioned in the meeting • who’s going to maintain this after handover That’s where architect and BA thinking shows up. Not in slides. Not in fancy reports/dashboards. But in decisions made before the first visual is created. The image below is a report I built early in my Power BI days (7 Years back :) ) when we didn't had much customizations available in Power BI. Simple layout, minimal visuals, focused on clarity over decoration. The best devs I’ve worked with: – challenge vague requirements – design the model like they won’t be around to fix it – care more about trust than colours – can explain a number without opening Power BI That mindset is what separates “dashboard delivery” from real impact. If this resonates, you’re probably my kind of audience. #PowerBI #MicrosoftFabric #DataAnalytics #AnalyticsCareers #BusinessIntelligence #DataProfessionals #News #Trending
14 Comments
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More