About
Yossi Matias is Vice President, Google, and the Head of Google Research. He leads…
Activity
62K followers
-
Yossi Matias reposted thisYossi Matias reposted thisLots of discussion out there about our next model, so I wanted to give an early look as soon as possible. Introducing Gemini 4 Argon! It shows frontier performance in complex workflows, cyber defense and software engineering. Teams are using it extensively at Google, from coding to quantum computing, great feedback. Importantly Argon has frontier safeguards and we are rolling it out responsibly - it’s with the US gov’t and going to a set of trusted cyber defenders through our Fairwind Program today. We’re going to make it available as soon as we can and as safely as we can. So hold tight, lots more coming, and you’re going to see us iterating rapidly. https://lnkd.in/gTiQb7DA
-
Yossi Matias shared thisLoved tuning into this Latent Space episode featuring our very own John Platt. In this in-depth conversation, John, a prolific pioneer at the intersection of AI and science, discusses some of the exciting work he is leading at Google Research, including how we can accelerate scientific discovery with Empirical Research Assistance (ERA), a novel approach for early fire detection with FireSat, AI based mitigation of climate impact by Contrails, Fusion and more (including how he discovered 2 asteroids, got an Oscar, and took a class with Feynman). Excited to work with John, and the strong Applied Science team he’s built - including key leaders like Michael Brenner, and Chris Van Arsdale. Thank you John!Yossi Matias shared thisI was a guest on the Latent Space AI Science podcast: https://lnkd.in/geuXrVtv . It was tremendously fun spending a couple of hours wandering through AI+Science space. Thanks for having me Brandon Anderson and RJ Honicky !🔬 An Oscar, Two Asteroids, and the Algorithm in Your sklearn: John Platt on AI for Science🔬 An Oscar, Two Asteroids, and the Algorithm in Your sklearn: John Platt on AI for Science
-
Yossi Matias shared thisA pleasure to launch ✨ The Google Research Podcast series, hosted by Sinead Bovell. This series will feature scientists and researchers discussing breakthroughs we are working on: from foundational machine learning, algorithms, and computing systems to AI for societal impact in health, education and planetary intelligence; to advancements in generative AI to quantum computing, and accelerating scientific discovery. In our opening episode, I sat down with Sinead to discuss some of our foundational research breakthroughs, often making what may have appeared impossible, possible. And how we are driving research breakthroughs in collaboration with many through the magic cycle of research to real-world impact on products, science, and society. We are excited about working towards AI as an amplifier of human ingenuity. And we believe that our ambitions for the societal benefits of AI must always be accompanied by an equal commitment to responsibility. Thanks to the incredible Google Research teams for all the impactful work (well more than can be covered in the series!). The work reflects deep collaboration with teams across Google and partners through academia and beyond. I hope you enjoy this discussion: https://lnkd.in/ghB9Vdng
-
Yossi Matias shared thisLearning is not a spectator sport. From foundational educational theories to modern cognitive research, it is well established that students learn best through active engagement and doing. Yet, digital learning often remains a passive experience. To help close this gap and empower educators, we are sharing our latest research on interactive learning. ✨ Generative UI for Education: We advance generative user interfaces (GenUI) for deeper educational journeys, allowing teachers to dynamically create custom, guided simulations tailored to their curriculum. ✨ Pedagogical Guardrails & Game Design: Drawing on learning science principles and the foundation of LearnLM, each simulation breaks complex topics into progressively difficult challenges complete with tailored feedback, scaffolded hints, and a built-in toolbox. ✨ Rigorous Self-Correcting Loops: We built agentic auto-evaluation processes into the creation pipeline—such as opening a Chrome instance to test solvability and adversarial actions—ensuring adherence to strict pedagogical and technical criteria. ✨ Working with Teachers: Initial studies and testing with expert STEM teachers show strong positive feedback. ✨ Expanding Pilot Program: We are releasing an initial sample library of some 30 AI-generated, teacher-reviewed STEM learning interactives, and schools using Google Workspace for Education can now sign up via the Google for Education Pilot Program. This research builds directly upon our efforts in education, following the 2024 release of LearnLM and the 2025 Learn Your Way research experiment that reimagined textbooks with generative AI. By advancing generative technologies for education, we are making progress towards a future where practice is active, effective, and tailored for every learner. Read more about our research and explore the library: https://lnkd.in/grhs6PAB Tech report: https://lnkd.in/gfuHynTg
-
Yossi Matias shared thisAs we expand our use of AI to address some of society’s biggest opportunities, one of the special areas is AI language research. Google’s journey in AI language research spans over 20 years. When we launched Google Translate in 2006, our goal was simple: to break down barriers between languages. Today we reached a significant milestone that stands as a testament to decades of AI research and advancement: Google technologies now support more than 300 languages, spoken by 7 billion people — representing 86% of the global population. Over the years I was passionate about how we can use technology to enable effective language understanding and communication and removing barriers of modality and languages. From translation and supporting numerous languages in Search experiences to Google Duplex, Live Caption and Read Aloud; from speech data gathering like Waxal, to incorporating cultural context and local nuances and cross lingual knowledge transfer in LLMs. Beyond translation and text processing, technology must understand how people actually communicate in the real world with all its cultural nuance and richness. 🗣️ From text-only to including tone, pacing, emotion, and context. By moving to native audio intelligence, systems like Gemini 3.5 Live Translate and Gemini 3.5 Transcribe process audio directly to naturally capture real-world communication phenomena like code-switching and emotional cues. 🗣️ From a few dominant languages to representing underrepresented and living languages accurately. Teaching AI to understand underrepresented languages required to rethink how we gather data. We rely on local grassroots partnerships like WAXAL in Sub-Saharan Africa and Project Vaani in India, gathering high-quality speech data that captures true conversational rhythms. And we teamed up with local experts, capturing local nuances with Amplify initiative. 🗣️ Overcoming real-world constraints: Access must be reliable where people live. Lightweight open models like TranslateGemma run efficiently on-device without cloud connectivity, while voice AI assistants like AVA bring Gemini to standard feature phones in low-resource regions. We’re also building on our work prioritizing open-source language innovation through our new tool Language Explorer. It’s an interactive tool that visualizes LinguaMeta, the world’s largest open-source language data repository, continuously mapping more than 7,000 languages. 🗣️ Designing for accessibility: We are expanding expression through initiatives like Sign Language-to-Text (SL2T) to support communities that rely on sign language to communicate. We'll continue to work closely with local communities to build technology that helps more people communicate, participate, and be understood on their own terms. AI as an amplifier of human ingenuity ✨ More on this work across Google in the blog by James Manyika: https://lnkd.in/euunWFZA
-
Yossi Matias shared thisFrom forecasting floods to tracking wildfires, Google is building state-of-the-art AI models and global partnerships — working toward the vision: no one should be surprised by a natural disaster.🌍 Today, we are proud to share our new white paper, “How Google AI is reshaping crisis resilience,” which outlines how we’re harnessing AI to predict disasters earlier, extend warning horizons, and empower communities before crisis strikes. In 2025 alone, we helped connect people with vital crisis information over 10 million times per day on average. Here are a few highlights of how AI is making a real-world impact: 🌊 Riverine Floods: Forecasting flood events up to 7 days in advance across ~150 countries, reaching over 2 billion people. ⚡ Urban Flash Floods: Predicting localized flash floods up to 24 hours in advance using Groundsource powered by Gemini. 🌀 Cyclones: Generating up to 1,000 predictive scenarios 15 days out with WeatherNext AI, providing forecasters an extra day of critical lead time. 🔥 Wildfires: Tracking active fire perimeters across 34 countries and co-developing FireSat to detect fires as small as a shipping container (5×5m) every 20 minutes. 📱 Earthquakes: Detecting seismic waves through our global Android network to deliver early alerts within seconds. 🤝 Local Action: Partnering with groups like UN OCHA, WMO, and GiveDirectly to translate early warnings into anticipatory aid, such as direct cash transfers before floodwaters hit. There is still much more work to do, but AI is proving to be an essential tool for climate adaptation and community resilience. 📖 Read the full white paper to explore our roadmap for AI-driven crisis resilience: https://goo.gle/3Uuzu6U
-
Yossi Matias shared thisJust published a viewpoint in Nature Medicine, sharing critical lessons learned from moving research with AMIE (Articulate Medical Intelligence Explorer) beyond in-silico simulations and patient-actor studies into real-world prospective clinical trials⚕️. Prospective evidence for conversational medical AI is hard, but non-negotiable. Together with our partners at Beth Israel Deaconess Medical Center, Harvard Medical School, Included Health and Stanford University School of Medicine we explore several foundational lessons: --> Trust is dynamic, not static: Trust cannot be captured by a benchmark metric. Real-world trust is earned incrementally through transparent, direct experience and longitudinal scientific evaluation. --> Safety is achievable with the right infrastructure: Across all 100 patient encounters at BIDMC under real-time physician supervision, safety supervisors did not need to intervene once. Rigorous cross-disciplinary review, clinical red-teaming, and conservative guardrails make safe evaluation possible. --> AI reshapes the clinical encounter: Primary care clinicians noted a meaningful shift from “data gathering” to “data verification.” Patients arrived prepared with coherent narratives—often feeling comfortable disclosing detailed histories in a pressure-free environment—freeing clinicians to focus on shared decision-making. --> Workflow integration matters: The promise of "triadic care" (patient–clinician–AI) only works when the handoff is seamless. If AI outputs do not integrate well with clinician routines and expectations, handoff friction can negate the benefits. Thank you to co-lead authors Mike Schäkermann and Cameron Po-Hsuan Chen, alongside our colleagues across Google Research, Google DeepMind, Google for Health, Beth Israel Deaconess Medical Center, Harvard Medical School, Included Health and Stanford University School of Medicine for this collaborative effort. Read more in Nature Medicine: https://lnkd.in/ga8FiSbT Full-text link: https://lnkd.in/gVuVUbZ6 BIDMC study: https://lnkd.in/dCPdebAJ Included Health study: https://lnkd.in/gRS94B5D
-
Yossi Matias shared thisMore updates on our ✈️ Contrails project: We are expanding our partnership with Cathay Pacific—our first commercial airline partner in Asia—to test and scale our AI-powered contrail mitigation technology across ultra-long-haul and transpacific routes. Here is how this progression works in practice: ➡️ Contrail mitigation represents one of the most immediately scalable and cost-effective climate solutions available today, working with existing aircraft without requiring new propulsion infrastructure. ➡️ Our system integrates AI predictions, satellite imagery, and weather intelligence to identify cold, humid atmospheric zones where warming contrails form. ➡️ Through Cathay Pacific's proprietary Electronic Flight Folder (EFF) connected via in-flight Wi-Fi, pilots receive dynamic forecasts directly in the cockpit to execute minor, safe altitude adjustments—much like navigating around turbulence. ➡️ Building on our earlier trials, an initial operational phase of over 80 flights achieved an estimated 40% reduction in contrail warming impact, with interventions on the Hong Kong–Singapore corridor alone accounting for over 50% of total emissions reductions. We are now scaling this into a second phase of trials with Cathay Pacific alongside Contrails.org to evaluate operational feasibility across diverse operating environments. By advancing open research together, we can accelerate progress in contrail science and unlock scalable climate solutions for aviation. More in the blog: https://lnkd.in/gKxFAZAx
-
Yossi Matias shared thisUnderstanding the genome is like understanding the language of life. Mastering it is a grand challenge that could transform our ability to understand biology and treat disease. New AlphaGenome Atlas by the Google DeepMind team is a platform containing predictions for the effects of 9 billion single-nucleotide variants (every single-letter change possible) in the human genome. Looking forward to seeing how researchers around the world will use it for understanding unsolved rare diseases, mapping rare variants associated with protein levels and complex traits and in general for accelerating genomics discovery! Read more below from my colleague Pushmeet Kohli.Yossi Matias shared thisToday, our team Google DeepMind is taking another step on our mission of deciphering the genome. We are releasing AlphaGenome Atlas, a massive (petabyte-scale) resource containing AlphaGenome predictions for every possible single-letter DNA change in the human genome — 9 billion in total. To help researchers navigate this data, we are introducing a new feature, the AlphaGenome Variant Impact (AVI) score, which combines predictions from our models, AlphaMissense and AlphaGenome, into a single number. This will help researchers quickly prioritize high-impact variants and better understand how they may be linked to disease mechanisms. Following in the footsteps of the AlphaFold Database, AlphaGenome Atlas aims to democratize the knowledge of variant effect prediction to researchers across the world and is available through a free, web portal, which can be used for any non-commercial research. From our early access program, we have already seen scientists successfully using Atlas to make progress on research related to understanding rare diseases and for isolating the variants responsible for disease phenotypes. This release marks another significant step towards our mission of deciphering the fundamental language of life. AlphaGenome Atlas: https://lnkd.in/ehhh3Hhj Understanding Life at Every Scale: https://lnkd.in/eBujddz7AlphaGenome Atlas: Molecular predictions for 9 Billion human DNA variantsAlphaGenome Atlas: Molecular predictions for 9 Billion human DNA variants
Yossi Matias commented on a post
1w
Thanks for the conversation, Sinead!
-
Yossi Matias liked thisYossi Matias liked thisWhy does the Google homepage still look so familiar nearly 30 years later, when Search itself has completely changed? That was one of the first questions Raúl Ordóñez asked me when we sat down in Madrid last week. The short answer: we strive to balance the simplicity people love about Search, while also massively upgrading the experience. Often easier said than done! We had a great conversation unpacking this and more, including how Search works today and where it’s headed next: 🔹 Moving beyond keywords so you can ask any question on your mind, no matter how specific or nuanced 🔹 How people are searching in totally new ways with voice, camera, and conversational follow-ups in AI Mode 🔹 Evolution to more personalized help with the ability to take action right from Search 🔹 How we’re building AI in Search to connect people with websites and creators across the web Thanks for the thoughtful questions, Raúl — I’ll keep practicing my Spanish for next time! :) Check out our full interview here: https://lnkd.in/gFUNgvJQ
-
Yossi Matias liked thisYossi Matias liked thisWhen I was thirteen, a clerk followed my grandmother and me through the Family Dollar store in our hometown in South Carolina, watching us as though we had come to steal something. I wanted to leave. Grandma Essie kept shopping. “Why would we leave?” she asked me. “We came here to shop, so we're gonna shop.” It was only years later that I began to understand how deliberately she had chosen her response. My grandmother’s example still shapes how I exercise judgment. Colleagues and leaders have taught me other lessons by trusting me with something new, challenging my thinking, or helping me see what I needed to change. As I mark a milestone at Google, I've been revisiting experiences like this with a question in mind: what is it like to work for me? In this article, I share a few of those experiences, including what it felt like to walk away from the identity I'd built as a lawyer, and how much energy I once spent trying to prove I belonged. These lessons continue to shape the leader I try to be. The people on my team are better placed than I am to say how often I live up to that. With thanks to Jacob Glick, Halimah DeLaine Prado and Karan Bhatia for their guidance.What Is It Like to Work for Me? The experiences and people that continue to shape how I lead.What Is It Like to Work for Me? The experiences and people that continue to shape how I lead.
-
Yossi Matias liked thisYossi Matias liked thisLots of discussion out there about our next model, so I wanted to give an early look as soon as possible. Introducing Gemini 4 Argon! It shows frontier performance in complex workflows, cyber defense and software engineering. Teams are using it extensively at Google, from coding to quantum computing, great feedback. Importantly Argon has frontier safeguards and we are rolling it out responsibly - it’s with the US gov’t and going to a set of trusted cyber defenders through our Fairwind Program today. We’re going to make it available as soon as we can and as safely as we can. So hold tight, lots more coming, and you’re going to see us iterating rapidly. https://lnkd.in/gTiQb7DA
-
Yossi Matias liked thisYossi Matias liked this23 years ago today, I officially started working for Google. Little did I know of the adventures that lay ahead, as I introduced myself to the Googlers sharing a small, rented office in London. If I could sit down and share advice with that younger version of myself, or anyone starting their career journey, there is a lot I would say. 1. Find your voice. Speak before you are “ready.” Don't wait for the perfect moment or for someone to ask for your opinion. I used to tell myself to just speak up at least once per meeting. When you have a seat at the table, use your voice, we need it. 2. Find your purpose. If you leave your family to come to work every day, it needs to be worth it. Quit if the passion is gone. If you show up, show up. Your impact matters, make it count. 3. Find your people. People that get you. Challenge you. Rely on you. Make you laugh. You are going to spend a lot of hours together. 4. Trust your gut. I’ve made a few decisions I regret because I let other people influence me. Your instincts are usually right, and will get you further than you think. 5. Treat your career like a jungle gym, not a ladder. Put your hand up for roles, ask for challenges, ask to be considered. There will be years where you feel like you’re flying and years where you’re just trying to keep your head above water. That’s okay. 6. What makes you different is your superpower. Growing up, I was one of only two girls in my advanced math class. Later on, I was an Irish woman navigating two male-dominated industries that had a very specific, established way of talking and acting. It took me a while to realize that my unique perspective was exactly why I was in the room in the first place. Don’t sand down your edges. 7. You can be "Mom-in-Chief." Before she passed, my dear friend Susan Wojcicki taught me that it’s possible to be both a good parent and a good leader, and that your kids will be proud of you for working on something that gives you purpose. 8. Always stay curious, learning keeps you young! To my team and all the amazing folks I work with every day, thank you for making the last 23 years the privilege of a lifetime. The 2003 version of me couldn't possibly imagine what we'd build together, and the 2026 version of me is incredibly excited for what's next.
-
Yossi Matias liked thisYossi Matias liked thisWow, summer is over already! This summer, I completed a Master of Public Health at the Keck School of Medicine of the University of Southern California. The program was eye-opening in so many ways, taking me through the epidemiology of chronic disease, environmental exposures, and infectious disease, along with health education, biostatistics, and the history of the U.S. healthcare system. I’m now armed with enough public health facts that I’m no longer invited to any parties! ;) More seriously, I’m incredibly grateful to the wonderful faculty, staff, and students at USC. I had the chance to dig into a wide range of topics, including: - The impact of prenatal care on preterm birth in the Medi-Cal community - The health effects of air pollution from the shrinking Salton Sea - The rapid increase in microplastic exposure and mitigation strategies ...and many more. A special thank you to Mellissa Withers, Ph.D., M.H.S., Jane Steinberg, PhD, MPH, Victoria Cortessis, Tracy Bastain, Steven Gazal, Tyler Mason PhD, Lee Eunjung, Sue Ingles, Lu Zhang, Sue Kim, Naying Z., and to my student colleagues. I genuinely enjoyed every class and hope our paths cross again soon. Thank you to Michael Howell, MD MPH and Karen DeSalvo for the encouragement along the way. For now, I’m continuing some of this work at Stanford’s March of Dimes Prematurity Research Center, working with Gary Shaw and team on research into preterm birth and health disparities. Time to put on a fall flannel ... and get those updated seasonal shots!
-
Yossi Matias liked thisYossi Matias liked thisThat's a wrap on the INSEAD AI Forum Americas 2026! Led by phenomenal speakers at the forefront of AI research and adoption, more than 200 participants took a deep dive into “AI's Next Frontier: Agents, Robots, and the Physical AI Economy” at the San Francisco Hub for Business Innovation. The takeaway from two days of conversations? Those who stay curious and humble will extract real value from AI. Hear it from: → Yossi Matias, VP of Google and GM of Google Research, sketched out Google's AI work across disaster prediction, healthcare, education, and science, as well the "magic cycle of research" with Heinz Blennemann (Principal, Blennemann Family Investments) → Martin Gonzalez(Organizational Design and Development Lead, Google DeepMind) shared Google's experiments with knowledge-sharing tools, while Maziar Brumand (VP of Product, ŌURA) explained how "tiger teams" help his healthwear firm innovate with AI within a constrained system. INSEAD Assistant Professor of Organisational Behaviour Michael Y. Lee moderated the lively panel → Kevin Schulman (Professor of Medicine and Professor of Operations, Information and Technology at Stanford University), Jiaxin Pei(Assistant Professor at the University of Texas at Austin, Fellow at Stanford HAI) and INSEAD's Assistant Professor of Technology & Operations Management Xinyu Liang, explored the design of AI workflow → Shweta Shrivastava(VP of Product Management at Waymo) and Vidya Sarma(Product Management Lead at Waymo) examined safety, privacy and the challenge of staying ahead in autonomous driving → Pinar Yildirim(Associate Professor of Marketing & Economics, The Wharton School), walked the audience through research on how robotization and AI affect human careers → Carlos Oliveros Forero (Head of Product, Robot.com) presented on the challenges and progress of deploying robots in the real world And the fitting finale was delivered by Vivienne Ming, Chief Scientist of Possibility Sciences. The author of the new book Robot Proof: When Machines Have All the Answers, Build Better People impressed the audience with her take on human-AI collaboration, and why traits like curiosity and intellectual humility matter for getting real value from AI. Thank you to our speakers, the INSEAD organizing team, and everyone who joined us in this important conversation. And there’s more to come! Mark your calendar for upcoming INSEAD AI Forums: → AI Forum Asia, Singapore, 30-31 October 2026 → AI Forum Middle East, Abu Dhabi, 27-28 January 2027 Full program of INSEAD AI Forum Americas: https://lnkd.in/dEj4fM5C Lily Fang Clarissa Taveras Alexandra Hansen Mark Caldwell Lauren Anderson, M.Ed. Chelsea Choppy Francisco Veloso
-
Yossi Matias reacted on thisYossi Matias reacted on thisBack in February, I sent Futurist Sinead Bovell an email pitching an idea for what we now call the Google Research Podcast. I’m so glad she said yes :) After many months of work behind-the-scenes (and a lot of time in frame.io), I’m excited to share our debut episode! Sinead sat down with VP & GM of Google Research, Yossi Matias, to unpack what happens when everyone gets a polymath in their pocket. They dive into: • Generative UI: Models generating personalized user experiences on the fly. • Accelerated Discovery: Turning months of literature review and hypothesis testing into days. • Ambient Intelligence: How AI is becoming as seamless and foundational as electricity. Building a show from scratch truly takes a village—lucky to have an amazing one: • Sinead Bovell & Yossi Matias for turning futuristic ideas into an engaging, grounded conversation. • Our truly tireless video production team for bringing polish and creativity to every frame. • Sinead’s team & our cross-functional partners spanning Research, Comms, Marketing, and Ops—we couldn't have done this without you. 🎬 Watch the full episode (link in comments!) and stay tuned for more conversations coming soon! Special thanks to Ronit Levavi Morad, Liat Ben-Rafael, Nick Singh, Allison Cullington, Kate Mischaikow, James Mulcahy, Kevin Soares, Cassandra Corbett, Maureen Fitzgerald, Jennifer Haslip, Brian Gabriel Jr., Isidora De Vicente, Marina Santalices Amigo, Tim Herrmann, Anastasia Komarova, Paul Plumeri Jr., Olivia Harris, Olivia Hoeft, Taylor Montgomery, Ruben Davis, Jacob Cohen ...and even my parents for watching the whole thing end-to-end twice!
-
Yossi Matias liked thisYossi Matias liked thisI’m thrilled to announce that Bessemer Venture Partners has raised $5.75B in new capital to back audacious founders: $1.75B for early stage and $4B for growth. This is remarkable time to be in the tech industry, but also to be a consumer of technology. The future is being decided by today’s entrepreneurs and startups, and it is a great privilege to be a witness and participant in some of their endeavors. As I approach 20 years as a partner at Bessemer Venture Partners, I can assure the Israeli tech community that we will continue to be an active investor here across both sectors and stages. Markets change. Governments change. Sentiment changes. Our conviction in Israeli startups will not. Thanks to all our founders, friends and colleagues, and thanks to our LPs for the continued support!
Experience & Education
Publications
-
Over 200 publications ( https://dblp.org/pid/m/YossiMatias.html )
-
See publicationhttps://research.google/people/YossiMatias/
https://scholar.google.com/citations?user=gtO5G0EAAAAJ
Patents
Honors & Awards
-
Kanellakis Theory and Practice Award
ACM
For seminal work on the foundations of streaming algorithms and their application to large-scale data analytics.
https://awards.acm.org/award_winners/matias_4099198 -
Fellow
ACM
-
Godel Prize
ACM-EATCS
For the profound impact on the theory and practice of the analysis of data streams
Other similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content