Kevin Trevey
Toulouse, Occitanie, France
4 k abonnés
+ de 500 relations
4 k abonnés
+ de 500 relations
À propos
🧭 I help organizations bring Trustworthy AI into their safety-critical system…
Activité
4 k abonnés
-
Kevin Trevey a republié ceciKevin Trevey a republié ceciWe had the honor of receiving the Microsoft Intelligent Manufacturing Award (MIMA) in the “Innovate!” category – live at Hannover Messe. The recognition highlights our internally developed AUMOVIO AI Hub – not just as a concept, but as a scalable, enterprise-ready AI platform already driving real impact across the organization. From engineering and software development to purchasing and manufacturing, AI agents are supporting end-to-end workflows and helping teams tackle complexity more efficiently. Even more meaningful: after winning the overall award last year, this year’s recognition once again underlines our ambition to turn AI into measurable value. A special thank you to everyone who contributed – in particular Teo Geneau, Kevin Trevey, Ben Aigner and Jean-Charles Huard 👏 Read more in our press release: https://ow.ly/RNxZ50YOX5p #AUMOVIO #MIMA #ArtificialIntelligence #IndustrialInnovation #HannoverMesse
-
Kevin Trevey a republié ceciKevin Trevey a republié ceciProud to see the AUMOVIO AI Hub winning the MIMA Award by Microsoft and Roland Berger in the Innovate category. Not another concept, but a scalable, enterprise-ready AI setup - built on a multi-agent architecture, driving end-to-end workflows across disciplines from engineering to purchasing. Even stronger in context: after winning the overall award last year with AI for Requirements Engineering, this year’s Innovate award highlights the consistent ability of AUMOVIO teams to turn AI into real impact. Outstanding work by Teo Geneau, Kevin Trevey, Ben Aigner and Jean-Charles Huard. Also great to see Erbe Elektromedizin GmbH recognized in the SME category with “EvidenceStream” - a strong example of AI delivering measurable value in regulated environments. It was a pleasure supporting you as Jury Sponsor throughout the process. Patricia K. Haller Christian Erbe Nermin Salkic Markus Enderle
-
Kevin Trevey a partagé ceciAI Programs don't fail because of technology. Andreas Horn is pointing out a global issue that companies are facing today 🧭Kevin Trevey a partagé ceci𝗔𝗜 𝗽𝗿𝗼𝗴𝗿𝗮𝗺𝘀 𝗱𝗼𝗻'𝘁 𝗳𝗮𝗶𝗹 𝗯𝗲𝗰𝗮𝘂𝘀𝗲 𝗼𝗳 𝘁𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆. ⇣ They die in the org chart. The pattern is always the same. 𝗛𝗲𝗿𝗲 𝗮𝗿𝗲 𝘁𝗵𝗲 𝗳𝗼𝘂𝗿 𝗞𝗶𝗹𝗹𝗲𝗿𝘀 𝗼𝗳 𝗔𝗜 𝗥𝗢𝗜: ⬇️ ➜ 𝗙𝗿𝗮𝗴𝗺𝗲𝗻𝘁𝗲𝗱 𝗢𝘄𝗻𝗲𝗿𝘀𝗵𝗶𝗽: CAIO, CTO, CIO, COO - everyone has a stake, nobody has ACCOUNTABILITY. AI becomes a political football, not a business capability. ➜ 𝗦𝘁𝗿𝗮𝘁𝗲𝗴𝘆 𝗳𝗼𝗹𝗹𝗼𝘄𝘀 𝗦𝗽𝗲𝗻𝗱: Licenses get bought. Pilots get launched. Then, months later, someone asks: "Wait, what precise, measurable problem are we solving?" There is no clear strategy in place - neither for data nor for AI - and it remains unclear which problems are actually meant to be solved. ➜ 𝗗𝗮𝘁𝗮 𝗕𝗹𝗶𝗻𝗱𝗻𝗲𝘀𝘀: Every GenAI use case hits the same quality, access, and governance wall. The people who know how to fix the data are often the last ones invited to the strategy room. ➜ 𝗦𝗵𝗮𝗱𝗼𝘄 𝗘𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻: The most valuable, successful AI work is often a solo side project in an Excel file. No sponsorship. No budget. No scale path. 𝗧𝗵𝗲 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 𝘀𝘂𝗰𝗰𝗲𝗲𝗱𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗚𝗲𝗻𝗔𝗜 𝗮𝗿𝗲 𝗡𝗢𝗧 𝘂𝘀𝗶𝗻𝗴 "𝗯𝗲𝘁𝘁𝗲𝗿 𝗼𝗱𝗲𝗿 𝗮𝗹𝗴𝗼𝗿𝗶𝘁𝗵𝗺𝘀". 𝗧𝗵𝗲𝘆 𝗮𝗿𝗲 𝘀𝗶𝗺𝗽𝗹𝘆 𝗳𝗶𝘅𝗶𝗻𝗴 𝘁𝗵𝗲𝗶𝗿 𝗼𝗽𝗲𝗿𝗮𝘁𝗶𝗻𝗴 𝗺𝗼𝗱𝗲𝗹: ⬇️ ✅ 𝗖𝗹𝗲𝗮𝗿 𝗢𝘄𝗻𝗲𝗿𝘀𝗵𝗶𝗽: Single point of authority with budget and mandate. ✅ 𝗘𝗺𝗯𝗲𝗱𝗱𝗲𝗱 (𝗗𝗮𝘁𝗮)-𝗚𝗼𝘃𝗲𝗿𝗻𝗮𝗻𝗰𝗲: Not an afterthought, but the design principle from Day 1. ✅ 𝗕𝘂𝘀𝗶𝗻𝗲𝘀𝘀-𝗙𝗶𝗿𝘀𝘁 𝗨𝘀𝗲 𝗖𝗮𝘀𝗲𝘀: AI initiatives tied directly to measurable revenue, cost, or risk metrics. ✅ 𝗣𝗿𝗼𝗽𝗲𝗿 𝗥𝗲𝘀𝗼𝘂𝗿𝗰𝗶𝗻𝗴: Execution teams treated as mission-critical, not a side-hustle. AI maturity isn't about technology. It's about organizational readiness. If this org chart looks familiar, the problem isn't your AI strategy. It's probably your operating model. ⇣ 𝗘𝘃𝗲𝗿𝘆 𝘄𝗲𝗲𝗸 𝗜 𝗰𝘂𝗿𝗮𝘁𝗲 𝘁𝗵𝗲 𝗯𝗲𝘀𝘁 𝗔𝗜 𝗱𝗿𝗼𝗽𝘀 - 𝗹𝗲𝗰𝘁𝘂𝗿𝗲𝘀, 𝗴𝘂𝗶𝗱𝗲𝘀, 𝗵𝗮𝗻𝗱𝘀-𝗼𝗻 𝘀𝘁𝘂𝗳𝗳 - 𝗶𝗻 𝗺𝘆 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 𝗛𝘂𝗺𝗮𝗻 𝗶𝗻 𝘁𝗵𝗲 𝗟𝗼𝗼𝗽 [𝗙𝗿𝗲𝗲]: https://lnkd.in/dbf74Y9E
-
Kevin Trevey a partagé ceci🚨 Automotive engineers & safety experts, the future is here! Last week at Société des Ingénieurs de l'Automobile, I had the opportunity to unveil our Microsoft-awarded solution, AUMOVIO AI Hub, and share our latest research on LLMs for Safety Engineering (available in comments)👉 💡 AI is no longer “just a tool”. Engineers who get hands-on now will define tomorrow’s automotive safety, mobility, and innovation. But technology alone is not enough. For AI transformation to succeed, it must be embraced and supported at the organizational level, with leadership, strategy, and culture aligned 🔧🤖 🙌 Huge thanks to SIA for the inspiring discussions.
-
Kevin Trevey a republié ceciKevin Trevey a republié ceci𝗧𝗵𝗲 𝗮𝗰𝘁𝘂𝗮𝗹 𝗿𝗼𝗮𝗱𝗺𝗮𝗽 𝗳𝗼𝗿 𝟵𝟭% 𝗼𝗳 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗚𝗲𝗻𝗔𝗜 𝗽𝗿𝗼𝗷𝗲𝗰𝘁𝘀 𝗶𝘀 𝗮 𝗰𝗼𝗺𝗽𝗹𝗲𝘁𝗲 𝗺𝗲𝘀𝘀. ↓ This visualization captures why so many Proof of Concepts (PoCs) are dying before production. (I've been in this specific terminal loop many times at my day to work). The unsexy truth is that enterprise transfromation is complex and only works when people, processes, and technology are properly aligned. And getting exactly that right means, you don't get a perfect linear path. You get friction. 𝟭. 𝗧𝗵𝗲 𝘂𝗻𝘀𝗲𝘅𝘆 𝗽𝗿𝗲𝗿𝗲𝗾𝘂𝗶𝘀𝗶𝘁𝗲𝘀 → "Fixing data issues" isn't a weekend task. It's months of sanitizing the raw material for your RAG architecture. → Legal, compliance and security will question everything. You need an automated, explainable governance loop bolted on before day one. 𝟮. 𝗧𝗵𝗲 𝗺𝗼𝗱𝗲𝗹 𝗵𝗮𝗹𝗹𝘂𝗰𝗶𝗻𝗮𝘁𝗶𝗼𝗻 𝘁𝗲𝗿𝗺𝗶𝗻𝗮𝗹 → "Model works... sort of" is the builder’s common language. (A nice way of saying 'this thing just cited non-existent regulation'). You find an edge case, sand it down, then realize context rot at 100k+ tokens means stripping back the prompt engineering entirel. 𝟯. 𝗧𝗵𝗲 𝗼𝗿𝗴𝗮𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗳𝗿𝗶𝗰𝘁𝗶𝗼𝗻 𝗽𝗶𝗽𝗲𝗹𝗶𝗻𝗲 → Security review will stop progress. It takes 3 months. (Parenthetical: because it should). → You finally launch, and user training is a hasty afterthought. Then you wonder why the business team isn't ready or why adoption is zero. I believe the only winning strategy is to embrace the tangle. You build assuming the data is imperfect, the stakeholder requirement will shift, and legal is slow. And that's why the human-in-the-loop—to debug the debugging—is essential. Anyone else currently in one of these loops? What's the point where you usually get stuck? Brilliant visual by Clare Kitching. ⇣ 𝗜 𝘀𝗵𝗮𝗿𝗲 𝗺𝗼𝗿𝗲 𝗼𝗻 𝗵𝗼𝘄 𝘁𝗼 𝗺𝗮𝗸𝗲 𝗔𝗜 𝗽𝗿𝗼𝗷𝗲𝗰𝘁𝘀 𝘀𝘂𝗰𝗰𝗲𝘀𝘀𝗳𝘂𝗹 𝗵𝗲𝗿𝗲. 𝗦𝘂𝗯𝘀𝗰𝗿𝗶𝗯𝗲 𝗳𝗼𝗿 𝗺𝗼𝗿𝗲: https://lnkd.in/dbf74Y9E
-
Kevin Trevey a republié ceciKevin Trevey a republié ceciDriving industrial innovation with AI We’re proud to share that we have been recognized with the Microsoft Intelligent Manufacturing Award (MIMA) in the category “Innovate!”. At the core: our internally developed AUMOVIO AI Hub, bringing artificial intelligence directly into employees’ digital workplaces. What does that mean in practice? 🔸AI integrated into everyday tools, systems, and workflows 🔸Supporting tasks from contract analysis and coding to maintenance diagnostics 🔸Enabling employees to create and use AI agents for automation and decision support 🔸Already 1,500+ AI agents deployed across the organization The result: teams save time, handle complex tasks faster, and work more productively. From the initial idea to enterprise-wide platform in less than 1.5 years, the AI Hub is now used by more than 21,000 employees worldwide. This is just the beginning. We believe AI will fundamentally reshape how industrial organizations work – and we are building the foundation for that transformation. Read more about this recognition: https://ow.ly/MTF350Yy29X #AUMOVIO #MIMA #ArtificialIntelligence #IndustrialInnovation #MIMA2026
-
Kevin Trevey a partagé ceciProud to see our hard work at AUMOVIO recognized at the Microsoft Intelligent Manufacturing Award 2026. Great achievement and teamwork behind this project. Congratulations to all the winners driving AI and digital innovation in manufacturing. Teo Geneau Jean-Charles Huard Laurent Julien Tom Spielvogel Ben Aigner Maria FelbermaierKevin Trevey a partagé ceci🏆 Congratulations to the winners of the Microsoft Intelligent Manufacturing Award 2026! Every year, the Microsoft Intelligent Manufacturing Award (MIMA), jointly organized by Microsoft Germany and Roland Berger, honors innovative digital solutions for Operations. Under this year's motto 'Transforming Operations for the Intelligence Era', the winners showcase exceptional use of artificial intelligence to enhance efficiency, transparency, resilience, and sustainability. 🌟 Our overall winner this year is Krones. The team has developed AI-powered digital twins which serve real-time process optimization in beverage manufacturing redefining factory performance by achieving higher yield with fewer resources. 🏅 Our category winners are: 🔹 Sustainability: tesa - AI platform strengthens resilience and decarbonization 🔹 Disrupt: EssilorLuxottica - Digital Product Passport solution for eyewear 🔹 Scale: Tetra Pak - Tetra Pak ® Factory OS™ is unifying factory data to accelerate digital transformation at scale 🔹 Add value: Kongsberg Digital and Yara International - Intelligent digital twin - a unified, real-time workspace for industrial decision-making 🔹 Innovate: AUMOVIO - Virtual Intelligence Orchestrator 🔹 SME (new category): Erbe Group- AI-powered clinical evidence mining transforms regulatory intelligence 💡 Learn more about the award-winning use cases on our website: https://lnkd.in/d8dNWNdv #RolandBerger #MIMA2026 #DigitalTransformation
-
Kevin Trevey a republié ceciKevin Trevey a republié ceciThe real reason AI is failing inside companies (and nobody wants to say it)... Here’s what the org chart looks like in most companies “going all-in” on AI: - The CEO announces a bold AI vision and tells the CTO to make it happen. - The CTO nods, shifts priorities, and hands the CIO the budget to “become AI-first.” - The CIO gets excited and starts building a brand-new AI empire: A Chief of AI? Of course. An AI Center of Excellence? Mandatory. An AI Ethics Officer? Absolutely. And yet… you already know the punchline. Everyone is talking about AI. Almost no one is building with AI. Except Jack. The intern. The only person actually shipping anything. And then leaders wonder: “Why isn’t our AI strategy working?” The answer is simple. We don’t need more AI committees. We need more Jacks. #ai
-
Kevin Trevey a partagé ceci🚨 Using AI in Safety-Critical Development? Read This First! Banning LLMs won't work. Neither will hoping existing standards are enough. If you're a safety engineer exploring AI tools, or already using them, our latest paper from ERTS 2026 cuts through the hype to address what really matters: managing the risks LLMs introduce in safety-critical domains. What we cover: ✅ The Real Risks – Beyond hallucinations: data quality issues, non-determinism, context failures, output bias, and architectural limitations that standards like ISO 26262 or DO-178C don't yet address. ✅ Practical Mitigation Strategies – Not just organizational policies. We detail technical safeguards: alignment techniques (RLHF, Constitutional AI), input/output guardrails, hybrid AI-human workflows with verification protocols, and domain-specific validation benchmarks. ✅ Real-World Case Study – Our enterprise AI platform (VIO) in automotive safety engineering, including a MISRA compliance assistant. We share pilot study findings showing both productivity gains and critical failure modes like overconfidence and context gaps—plus how we mitigated them. The bottom line: Responsible LLM deployment in safety-critical domains requires measurable safeguards, preserved human expertise in critical decisions, and continuous validation frameworks, not generic AI guidelines. If you're navigating the intersection of AI and functional safety, this paper offers a framework grounded in real engineering practice. Authors : Hugues Bonnin Kevin Trevey Teo Geneau Jing Xiao #FunctionalSafety #AI #SafetyEngineering #ISO26262 #Automotive #LLM #ResponsibleAI #ERTS2026 #TrustworthyAI
Kevin Trevey a répondu à un commentaire
2 sem.
Carlos V. Roman most valuable comment in this thread ! 👍🏼💯
Kevin Trevey a commenté un post
2 sem.
What a journey! 🧭
We are proud to deliver our award-winning AI platform to more than 20k+ users at AUMOVIO
Kevin Trevey a commenté un post
3 sem.
The journey brought us here, now onto what’s next 🧭
A great team with bold ambition! Thanks Michael Sicker for your support!
-
Kevin Trevey a aimé ceciKevin Trevey a aimé ceci$180 in hardware. Two weekends. Zero lines of code. Claude built it. Tested it. Fixed its own failures. The drone tracks targets with a laser. Autonomously. No human wrote a single line. The AI just... figured it out. Now imagine what else it's figuring out. 🦞
-
Kevin Trevey a aimé ceciISO/IEC TR 5469 isn't the answer to functional safety for AI. It's the first credible framing of the question. That distinction matters. Functional safety standards were built on a deterministic worldview; specify behavior, implement it, verify implementation matches spec. Machine learning breaks every link in that chain. The behavior is learned, not specified. The implementation is a tensor of weights nobody can review. Verification against a specification is impossible because no specification exists in the traditional sense. TR 5469 doesn't solve this. It's a Technical Report, not an International Standard, the "TR" is load-bearing. You can't be "compliant" with it the way you can be compliant with ISO 26262. What it does give you: → A vocabulary shared between safety and ML communities (where before there was mutual incomprehension) → A classification framework (three usage classes, A/B properties) that's genuinely useful as a triage tool → A structure for the conversation with assessors → Honest acknowledgment of what it doesn't know I put together a practitioner's guidebook that walks through the framework, traces a worked AEB pedestrian-detection example end-to-end, and is explicit about the gaps, quantitative ASIL claims for ML, statistical evidence sufficiency, drift and retraining lifecycle, foundation models, online learning. ISO/IEC 22440 is coming and may turn parts of this informative guidance into shall-statements. #FunctionalSafety #ISO26262 #AISafety #AutomotiveSafety #SOTIF
-
Kevin Trevey a aimé ceciKevin Trevey a aimé ceciEveryone talks about Europe falling behind in AI. Meanwhile, it is already running on production lines. In this Industry Future Talk, Bernd Hagmanns from Krones shows how they turned a digital twin into an operational copilot that interacts with live production. But what struck me most was not the technology. It was how they built trust on the shop floor: not by telling engineers to believe in it, but by showing them the results and taking up their input until they trusted the output themselves. This is exactly what we see across industries: the technology is ready. The bottleneck is adoption. And adoption is a leadership challenge, not a technical one. And one clear takeaway on partnerships: when digitalization moves this fast, no company can do it alone. You need a network. Thanks Bernd for proving a point that matters far beyond Krones. The companies that will lead in industrial AI are not the ones with the best models. They are the ones where people on the shop floor trust the output. Roland Berger #IndustryFutureTalks #AI #Implementation
-
Kevin Trevey a aimé ceciKevin Trevey a aimé ceciAt the invitation of the works council, I attended today’s employee meeting in Frankfurt together with employee representatives and Kaweh Mansoori, Hesse’s Minister for Economic Affairs, Energy, Transport, Housing and Rural Areas (Hessisches Ministerium für Wirtschaft, Energie, Verkehr und Wohnen). The discussion focused on how we, as a company, provide direction and navigate responsibly through a phase of major transformation – both within the organization and with a view to Germany as a business and industrial location, and the respective roles of companies, social partners, and policymakers. Periods of transformation often involve complex structural questions for companies. Addressing them requires careful consideration and a responsible weighing of different options. Precisely because this situation involves uncertainty for many colleagues, dialog, social partnership, and shared responsibility are even more important – for the perspectives of our employees, for the continued development of AUMOVIO as an independent company, and for the future of Germany as an industrial location. #AUMOVIO #Leadership #HighPerformanceOrganization
-
Kevin Trevey a aimé ceciKevin Trevey a aimé ceciWe delivered a solid improvement in operating performance in the first quarter of 2026 – despite a persistently challenging market environment. Adjusted consolidated sales (Q1 2026): €4.4 billion Adjusted EBIT (Q1 2026): €106 million Adjusted EBIT margin (Q1 2026): 2.4% Overall, the results demonstrate our disciplined execution and consistent progress in a challenging environment. Our CEO Philipp von Hirschheydt: “In the first quarter, a persistently challenging market environment, combined with material currency effects and volume declines, affected our business performance. Nevertheless, we succeeded in further improving profitability – a clear indication of the effectiveness of our ongoing efficiency measures, which will continue to play a central role going forward.” Our CFO Dr. Jutta Doenges: “Strict cost and cash management is our top priority – and we continued to make further progress in the first quarter. The improved free cash flow and strong net liquidity give us the financial flexibility to navigate a challenging market environment with confidence, while at the same time laying the foundation for further profitable growth.” Read our press release for more information: https://ow.ly/zoip50YVT4a #AUMOVIO #QuarterlyResults #technology #innovation #mobility
-
Kevin Trevey a aimé ceciKevin Trevey a aimé ceciA coffee chat with the CEO. ☕ Last week, Philipp von Hirschheydt visited our North America headquarters in Auburn Hills, MI. One meeting during the visit included an informal coffee chat, creating an opportunity for open dialogue - allowing employees to share perspectives, ask questions, and gain insights into the business, global markets, our culture, and the automotive industry. Just one stop on a full agenda, but a great reminder of the power of conversation and connection. #AUMOVIO #Leadership #FutureMobility
-
Kevin Trevey a aimé ceciKevin Trevey a aimé ceciBehind every innovation, there are people. Progress doesn’t just happen – it’s driven by individuals who bring in ideas, take ownership and move things forward every day. We believe that creating meaningful impact starts with an environment where people stay ahead of the curve, are reliable no matter what, grow stronger together, and are committed to win. Thank you to everyone who contributes, collaborates, and helps build what’s next. #AUMOVIO #OwnWhatsNext
-
Kevin Trevey a aimé ceciKevin Trevey a aimé ceci🎥 Telling stories about innovation in manufacturing. It was inspiring to be back at Hannover Messe last week for the #MIMA2026 Award Ceremony, celebrating this year’s most outstanding use cases in operations and manufacturing. This marks the 7th edition of #MIMA (the Microsoft Intelligent Manufacturing Award) by @RolandBerger and @Microsoft, and my 6th year of making films about it and telling the stories of the winning use cases and their creators. A fascinating task. 🚀 MIMA keeps growing. The award has been extended to EMEA. The number of categories has increased again, and it now covers a wide range of industries, from automotive, the medical sector and the food & beverage industry to consumer goods, the energy field and beyond. With it, the variety of stories told has expanded too. What stood out to me: AI has become truly central to innovation. Companies now embed AI in their core operations, and the impact on resilience, competitiveness and sustainability is remarkable. For me that means, AI can be seen, it is visible on camera. A big thank you to Roland Berger, Microsoft and the fantastic MIMA team for their trust in all these years. 🙏 I am already excited about what’s next! 🏆 MIMA 2026 Winners Overall Winner : Krones (Germany) Innovate! : AUMOVIO (Germany) Scale! : Tetra Pak (Switzerland) Add Value! : Yara International & Kongsberg Digital (Norway) Disrupt! : EssilorLuxottica (France/Italy) Sustainability! : tesa (Germany) SME! : Erbe Group (Germany) The film about this year’s award ceremony at Hannover Messe will be online soon. #RolandBerger #MIMA2026 #MIMA #CorporateFilm #Storytelling #VisualStorytelling #Innovation #Manufacturing #AI #HM26
Expériences de bénévolat
-
Assistant Scrutineer for FIA
Formula 1
- 1 an 1 mois
French Grand Prix 2021 - Paul Ricard Circuit
French Grand Prix 2022 - Paul Ricard Circuit -
Projets
-
🚀 AUMOVIO AI Hub
From a side project to 1,800+ daily users
Autres créateurs -