Multiverse Computing

Great to be part of initiatives like RETECH: Q-AINA alongside Qilimanjaro Quantum Tech , Barcelona Supercomputing Center and Eurecat - Technology Centre 🚀 At Multiverse Computing, we strongly believe the future of AI will be shaped by hybrid approaches where quantum and quantum-inspired techniques enhance classical systems to make them more efficient, scalable, and sustainable ⚡🤖 Projects like Q-AINA are key to validating real-world applications, from optimizing language models to pushing the boundaries of AI performance with fewer resources 📊🌍 #AI #QuantumComputing #ModelCompression #DeepTech #Innovation

Yesterday our co-founder & CEO, Enrique Lizaso Olmos, took part in the discussion around the IN+DEF program a key initiative shaping Spain’s industrial and technological future through defense🛡️ At a time when public-private collaboration and innovation are more critical than ever, forums like this highlight the role of efficient AI in transforming the defense sector 🤖 At Multiverse Computing, we remain committed to delivering compressed AI models to help build a more efficient, resilient, and future-ready defense industry 🚀

As AI models scale, efficiency is no longer optional it is the defining challenge ⚡ Recent innovations like Google’s TurboQuant highlight how critical it is to optimize inference. At Multiverse Computing, we are tackling the same problem from a different angle with CompactifAI🔬 The key insight? These approaches are not competing they are complementary 🤝 🔹 CompactifAI reduces model size before deployment • Shrinks models by up to 90% 📉 • Cuts memory requirements dramatically 💾 • Enables deployment on smaller, more affordable hardware 💻 🔹 TurboQuant optimizes performance during inference • Reduces runtime memory usage 📊 • Speeds up attention computations ⚡ • Improves long-context efficiency 📚 Together, they unlock a powerful, layered approach to AI efficiency ➡️ Compress the model ➡️ Deploy on leaner infrastructure ➡️ Optimize runtime performance Because in reality, the biggest cost in AI is not just running models it is hosting them. At Multiverse Computing, we believe the future of AI efficiency lies in stacking innovations across the entire stack from architecture to runtime. 🔗 Read the full article: https://lnkd.in/eFPZw9ex #AI #MachineLearning #LLMs #Efficiency #DeepTech #CompactifAI

Are you ready to build the next era of AI? At Multiverse Computing, we are expanding our global footprint and looking for the boldest minds in the industry to join our mission. If you want to work at the intersection of deep tech and real-world impact, the challenge starts here. Open Positions: LLM Engineer | Donostia, Spain 🇪🇸 https://lnkd.in/emhPY94E Enterprise Account Executive | Poland & Eastern Europe 🇵🇱 https://lnkd.in/eFcAiQSC Strategic Account Executive (Telco & Hi-Tech) | UK 🇬🇧 https://lnkd.in/ecnzqCvb Enterprise Account Executive | California, US 🇺🇸 https://lnkd.in/exa_xxbY Strategic Account Executive (Public & Defense) | Toronto, Canada 🇨🇦 https://lnkd.in/ec2mZWb2 Not your perfect match? We have 35+ more opportunities waiting for you across engineering, sales, and operations. Explore all open roles: https://lnkd.in/g2R8c6K9

Last friday Multiverse Computing was recognized at the 2026 National Computer Engineering Awards (Premios Nacionales de Ingeniería Informática) 🏆 Our co-founder & CEO Enrique Lizaso Olmos together with co-founder & CMO Alfonso Rubio-Manzanares received the award in the spanish senate. Our AI compression solution CompactifAI received the award for Best Private Sector Project, highlighting its impact in making advanced AI models more efficient, scalable, and sustainable. This recognition celebrates the work of our entire team pushing the boundaries of AI using quantum-inspired techniques to dramatically reduce the cost and energy consumption of large models without compromising performance 📈🔋 Organized by Consejo General de Colegios Profesionales de Ingeniería Informática - CCII , these awards honor innovation and excellence in technology and its contribution to society. A huge thank you to everyone who made this possible and to our partners and clients who trust us to redefine what’s possible in AI 🤖🔥 #AI #ArtificialIntelligence #Innovation #DeepTech #MachineLearning #CompactifAI #MultiverseComputing

Great to see this collaboration taking shape alongside Telefónica Tech, Qilimanjaro Quantum Tech and QCentroid, bringing quantum computing closer to real-world AI applications. By integrating cutting-edge infrastructure with advanced software and expert services, we are helping organizations address complex challenges and enhance model performance. At Multiverse Computing, we believe in the power of strong ecosystems to bridge the gap between innovation and practical deployment. Excited to be part of this journey 🚀

Our CEO, Enrique Lizaso Olmos, has been honored at the prestigious Tech Tour Growth Europe 2026. This recognition solidifies our position at the very forefront of the European Deep Tech sector, highlighting our growth and the real-world impact of our technology. Together, we are pioneering the era of Efficient and Sovereign AI.   𝗥𝗲𝗮𝗱 𝘁𝗵𝗲 𝗳𝘂𝗹𝗹 𝗮𝗻𝗻𝗼𝘂𝗻𝗰𝗲𝗺𝗲𝗻𝘁: https://lnkd.in/dM73_Ws4

How do you prepare an autonomous system for hazards that aren't even in your training data? 🚗🔍 In autonomous driving, the "Cold Start" problem is a massive hurdle: you can’t possibly label every potential obstacle. In our latest blog post, Ahmet Gürhan (Machine Learning Engineer) explains how we flipped the script. By leveraging a self-supervised AutoEncoder, we shifted the focus from the "hazard" to the "road." The logic is simple: Instead of explicitly detecting anomalies, we model the features of a "normal" road surface; any deviation from this distribution is then flagged as an anomaly. Technical Highlights: ✅ Zero Labeling Costs: 100% self-supervised process leveraging the BDD100K dataset.  ✅ Edge-Ready: 83.3% model reduction (66M to 11M parameters) via proprietary HOSVD.  ✅ Real-Time Safety: < 0.3s latency, fully optimized for immediate on-road action. This project proves that efficiency and accuracy can go hand in hand without relying on massive infrastructure. Want to see the full architecture and our GMM scoring logic? 👇Find the link to the full blog post in the first comment below.

Proud to be part of the conversation on Europe’s technological sovereignty in defence 🇪🇺🛡️ At the Spanish pavilion coordinated by TEDAE, Asociación Española de Empresas Tecnológicas de Defensa, Seguridad, Aeronáutica y Espacio, Alberto Reguera and María Bertomeu Pardo had the opportunity to share how Multiverse Computing is helping advance AI capabilities for the defence sector through compressed AI models. 🤖⚛️ #Defence #AI #TechnologicalSovereignty #Europe