动态
1565 位关注者
-
Pierre Garrigues 转发了此动态Pierre Garrigues 转发了此动态The optimizations in this guide compound. Better fraud scoring means fewer fraudulent transactions reach authorization. Stronger authentication means more transactions carry liability shift. Post-authorization interventions reverse high-risk charges before they become disputes. We built these systems to be aware of their impact on every other stage, and the results show why that matters. New guide from the team: https://lnkd.in/gVavhW86Optimizing payments at scale: How Stripe applies AI across the payment lifecycle | StripeOptimizing payments at scale: How Stripe applies AI across the payment lifecycle | Stripe
-
Pierre Garrigues 转发了此动态Pierre Garrigues 转发了此动态Stripe handles billions of dollars in payments every day—and that number soars during major shopping periods. Over Black Friday and Cyber Monday 2024 alone, businesses processed more than $31 billion on Stripe. With volume spanning more than 10,000 issuing banks and hundreds of currencies, countries, card products, and payment features, performance issues can arise from unique combinations of these factors, like a spike in failures on digital wallet payments from debit cards issued in France. To keep all transactions running smoothly, our real-time ML system detects and fixes performance issues on specific subsets of payments traffic, ensuring payments go through without a hitch. More in our blog.Using ML to detect and respond to performance degradations in slices of Stripe paymentsUsing ML to detect and respond to performance degradations in slices of Stripe payments
-
Pierre Garrigues 分享了此动态This is a great opportunity to join a team at the forefront of ML innovation at Stripe, please get in touch with me or Swara if you’re interested!Pierre Garrigues 分享了此动态We're looking for a Machine Learning Engineer to join our team and drive innovation in payments technology. You will : • Shape long-term ML strategies for Stripe • Deliver business impact through ML solutions • Develop advanced ML models, including foundation models, sequence models, LLMs, and DNNs on tabular data Join us for the next challenge in your career! We're open to all locations in United States. Apply today or message me to learn more! #hiring #mlengineer #machinelearning #deeplearning #mljobsMachine Learning Engineer, Payments ML AcceleratorMachine Learning Engineer, Payments ML Accelerator
-
Pierre Garrigues 分享了此动态Come join us for an evening with our teams and get an inside look at the AI/ML initiatives that are happening at Stripe! Fill out the interest form now to let us know you want to attend: https://lnkd.in/g2ybDyHFPierre Garrigues 分享了此动态Want the inside scoop on AI/ML at Stripe? Join for this intimate event with our team in SF! We'll dive deep into our stack, how AI powers Stripe today, and what you could build with us next. Just fill out the interest form now to let us know you want to attend: https://lnkd.in/d6_xxqim #ArtificialIntelligence #MachineLearning #AI #ML Gautam Kedia Karan Sharma Vladimir Zhukov Stathis Vafeias Pierre Garrigues Inayat Khosla Swara Mehta Kumar Chellapilla Fabi Prabhakar Twum Djin
-
Pierre Garrigues 转发了此动态Pierre Garrigues 转发了此动态My team is hiring for a data analyst! I work in an area that's high priority, high impact and high growth. My partners understand what data science and data analytics can bring to the table, and are some of the most collaborative I've worked with at Stripe. Please apply through the website, or, if you know me personally, ping me for a referral!
-
Pierre Garrigues 转发了此动态Pierre Garrigues 转发了此动态Are you a current PhD or Masters student interested in working at Stripe? We're hiring Data Science and Machine Learning interns for Summer '24! Join us to make a tangible impact on the world's economic infrastructure and shape the future of global commerce. ‣PhD, Data Scientist: https://lnkd.in/gRYGyCqP ‣Masters, Data Analyst: https://lnkd.in/gsASuHPz ‣PhD, Machine Learning Engineer: https://lnkd.in/gtDgHfUZ
-
Pierre Garrigues 分享了此动态My team at Stripe is looking for talented ML Engineers and Data Scientists who want to join a newly formed team to accelerate our usage of ML across Stripe’s Payment products. This is an exciting opportunity to work on cutting edge AI/ML and advance Stripe’s mission to grow the GDP of the Internet! Please use the link below to apply and ping me or Swara Mehta if you are interested!Pierre Garrigues 分享了此动态I'm hiring exceptional machine learning engineers in US/Canada! This is an incredible opportunity to work on advanced ML applications. We are developing cutting edge deep learning models tailored to Stripe’s payment data, and infrastructure to enable rapid ML exploration. Apply directly or ping me if you'd like to learn more about the opportunity! https://lnkd.in/gYeEXjDT #hiring #mlengineer #machinelearning #deeplearningMachine Learning Engineer, Payments ML AcceleratorMachine Learning Engineer, Payments ML Accelerator
-
Pierre Garrigues 发布了此动态I’m happy to share that I’m starting a new position as Head of Core Payments Data Science at Stripe!
-
Pierre Garrigues 分享了此动态Pierre Garrigues 分享了此动态Heads up developers and engineers! We're very proud to announce that Vespa, Yahoo's big data processing and serving engine, is available as open source on GitHub. We're making it easy for anyone to build applications that can compute responses to user requests, over large datasets, at real time and at internet scale – capabilities that up until now, have been within reach of only a few large companies. Take a look, start building, and let us know what you think! (Great work by Jon Bratseth, Peter Cnudde, Gil Yehuda, and our whole team in Trondheim, Norway!)
-
Pierre Garrigues 赞了此动态Pierre Garrigues 赞了此动态Today, we're announcing $35 million in Series A funding led by New Enterprise Associates (NEA), with participation from Town Hall Ventures, SignalFire, and existing investors Moxxie Ventures, Uncork Capital, Coalition Operators, The Aligned Fund and Frist Cressey Ventures. But the opportunity we're pursuing is much bigger than any one funding milestone. Healthcare has an extraordinary amount of medical knowledge, specialized expertise, and increasingly sophisticated technology. Yet we still struggle to connect the right patients to the right capabilities at the right time. We founded @Knit to solve this problem. We're building a world model for healthcare. Our model, Med-BEAM, learns from billions of real clinical decisions and hundreds of millions of longitudinal patient records to understand how patients move through the healthcare system, what happens next, and how different decisions can change their trajectories. This is fundamentally different from simply teaching an AI model what the medical literature says. It's about understanding how healthcare actually works. And the results are striking. In benchmarks predicting real clinical events and care decisions, Med-BEAM was: 🟢Up to 4x more accurate than leading language models 🟢100x faster 🟢250x less expensive This creates an entirely new infrastructure layer for healthcare: clinical intelligence that can power everything from patient navigation and specialist routing to hospital operations and the next generation of AI agents. I'm incredibly proud of the Knit team building Med-BEAM, including our investors and early customers. I wrote more about why we founded Knit, the science behind our approach, and what we believe this makes possible for healthcare: https://lnkd.in/gmjVMEaf #WorldModels #HealthcareAI #ClinicalIntelligence
-
Pierre Garrigues 赞了此动态Pierre Garrigues 赞了此动态Trying a new approach to the working-parent juggle: bringing one of my kids as my plus-one when I can. My job involves a lot of dinners and events with customers and partners. It’s one of my favorite parts of the role, but it also means a lot of evenings away from my family. This weekend I hosted a group at the Warriors game and brought my little guy along. I got to spend time with customers and partners, while also having a pretty special night with my son. It won’t work for every event, and I know I’m lucky to have the flexibility — grateful to work somewhere that supports it 🙏
-
Pierre Garrigues 赞了此动态Pierre Garrigues 赞了此动态🇬🇧 I am honored to be featured in l'Opinion "100 shaping AI in Europe" list. This is a testimony of the work of the Google DeepMind and Google Research teams that push the boundaries of what we can achieve with AI every day, especially in healthcare. I am proud to share this journey with other architects and pioneers building a strong, competitive, and responsible AI ecosystem in Europe. 🇫🇷 Très honorée de figurer dans le palmarès des « 100 qui font l'IA en Europe » réalisé par l'Opinion. Cette nomination rend hommage au travail des équipes de Google DeepMind et Google Research, qui repoussent chaque jour les limites de ce que l'on peut construire avec l'IA, en particulier en santé. Bravo aussi à tous les architectes et pionniers qui contribuent à développer un écosystème de l'IA fort, responsable et compétitif en Europe.
-
Pierre Garrigues 赞了此动态Pierre Garrigues 赞了此动态Are you an ambitious, technical, AI-pilled Data Science manager? Have you thought about how you could help increase the GDP of the internet via data, ML, and AI? Could you imagine a more engaging dataset than transactions representing 1.73181567% [0] (and counting) of global GDP? We're hiring across the (extremely talented) team, but especially looking for experienced and technical managers. Reach out if you're interested... and if you aren't interested we think you will be after reading our dev blog: https://stripe.dev/ :) [0] https://stripe.com/
-
Pierre Garrigues 赞了此动态Pierre Garrigues 赞了此动态Last week we took the team out for dinner. It is always great to spend some time together outside the office, away from the usual topics and conversations we have during work. Especially in Berlin, with so many different cuisines, it’s fun to try out new restaurants together. A perfect combination of good food and quality time with the team. We went to a ramen & natural wine restaurant and found out that most of the team actually hates natural wine 😂. At least the ramen was amazing, good choice Caterina!
-
Pierre Garrigues 赞了此动态Pierre Garrigues 赞了此动态I often get asked whether fine-tuning LLMs is worth it. We find it generally is when you need precise evaluations combined with the flexibility to use rich, heterogeneous evidence. Take risk decisioning at Stripe. LLMs are promising here because they understand how real businesses work and can reason over payments, text, and time-series data serialized into a single record. Out of the box, though, they're word predictors, not built for the rigorous statistical analysis our risk decisions need. On a merchant fraud-review cohort, zero-shot prompting produced 0.49 precision at human-review recall, domain-guided prompting 0.71, and human reviewers 0.79. So we fine-tuned an open-weight 4B model on review labels and read a continuous fraud score directly from its yes/no token probabilities. Average precision rose 24% over domain-guided prompting, and the model outperformed human reviewers at the same recall. For those curious about the mechanism, linear probes suggest an important driver of the boost is that fine-tuning made the signals of *legitimate* businesses easier to decode. Full write-up at https://lnkd.in/gY5PQFmq
-
Pierre Garrigues 赞了此动态Pierre Garrigues 赞了此动态AI is likely the biggest driver of change most of us will see in our lifetimes. And yet many takes on how it will shape the economy aren't yet grounded in data. We want to help change that. Stripe sits on one of the richest real-time datasets for understanding how AI is reshaping the economy. Our latest post lays out our empirical agenda for AI. It covers the questions we think matter most and where Stripe data can help answer them. We're publishing it to show what we're thinking about and to find people working on the same questions. If that's you, we'd love to hear your ideas and proposals for collaboration at stripeeconomics@stripe.com. More here: https://lnkd.in/g8KMVusv
出版作品
-
Group Sparse Coding with a Laplacian Scale Mixture Prior
Advances in Neural Information Processing Systems 23 (NIPS)
-
An Homotopy Algorithm for the Lasso with Online Observations
Advances in Neural Information Processing Systems 21 (NIPS)
-
Learning Horizontal Connections in a Sparse Coding Model of Natural Images
Advances in Neural Information Processing Systems 20 (NIPS)
专利发明
-
Automatic Image Piling
颁发日期: US US20140254946
查看专利A system for grouping together visually similar images for photo navigation, organization and discovery.
收到的推荐信
1 位会员推荐了Pierre
加入领英,即可查看查看Pierre的完整档案
-
浏览共同好友
-
请求引荐
-
直接联系Pierre
其他相似会员
浏览更多动态
-
Olivier Le Douarin
KoBold Metals • 1505 位关注者
My time at Stripe has come to an end after 4.5 incredible years. Stripe is truly a very special place, and the exposure to a multitude of businesses, the pace of innovation, the contagious curiosity of so many peers, all inspired me to pursue a new opportunity that my inner engineer couldn't pass up. I’m excited to share that I've joined KoBold Metals, where I will be diving into critical mineral exploration, blending AI and Data Science with the physical world to help enable the clean energy transition!
190
31 条评论 -
Anthony Voellm
Stripe • 4105 位关注者
In a recent mentoring discussion, I reflected on my past fears - the fear of running out of ideas. This feeling would be even more intense today, especially with the rise of AI coding tools. However, I realized long ago that I will never truly run out of ideas—only the time to implement them. As we embrace today's AI coding tools, we are entering an era where ideas will be the primary determinant of success, rather than coding skills. AI can code... you need to bring the ideas. This fundamental shift to ideas bringing value is upon us. You have all the skill you need at your fingertips.
39
1 条评论 -
Claire Vo
CXO.dev • 7万 位关注者
The team at Stripe is setting the standard for internal AI platforms: minion coding agents, a custom prototyping rig, and now their company brain, Kai. On today's episode of How I AI, Sharadh Krishnamurthy shows us how 1.5 engineers and 2 weeks got them a company brain, including: - projects as governance - skill routing + telemetry - a skills platform that works for 10k teammates Plus, he and I debate the merits of gentle parenting your AI (esp when your company is running evals.) Full episode on YT: https://lnkd.in/gdiKFtwe
175
23 条评论 -
Sen Zhuang
Stripe • 1148 位关注者
This one feels important to share! Subscriptions get complicated across currencies - between FX risk, conversion fees, and renewal surprises, there is a lot of friction. We’ve been working hard to bring Adaptive Pricing to subscriptions, allowing businesses to easily present customers in their local currency while Stripe handles the conversions and stability behind the scenes. We also rigorously validated the impact it has on Stripe users. The results? A 4.7% uplift in conversions, a 1.9% increase in authorizations, and 5.4% higher LTV! Higher conversions, fewer declines, and stronger LTV mean happier customers and more revenue for businesses globally. Learn more: https://lnkd.in/gaK92pvy
40
-
MotherDuck
3万 位关注者
He built AI models & production-grade agents at scale for Meta. Now he runs the advanced AI group for one of the world's leading financial platforms. Tosh Rayadhurgam is Head of Advanced AI at Stripe, where his group builds Stripe's foundational GenAI models and LLM applications. Before Stripe he led foundational AI efforts at Meta, with earlier stops at Lyft and Amazon Lab126. On Nov. 5, Tosh joins Sudeep Das of DoorDash and Tomasz Tunguz of Theory Ventures on the "AI at Scale Inside Real Companies" panel at Data Outpost. Tosh and Sudeep will get into how they evaluate models and make build-vs.-buy decisions once AI has to serve millions of users under latency and cost constraints. 🎟️ Tickets: dataoutpost.ai
15
3 条评论 -
Kevin Miller
Stripe • 2万 位关注者
What will it take for agentic payments to move from early adoption to everyday infrastructure? At MPP IRL, Dan Romero, Patrick Collison, and I heard some spicy takes from builders across a range of companies, from Replit to Amazon to Visa (and many more!) on exactly that question. Attendees got hands-on with the Machine Payments Protocol and gave us direct feedback: Where are the biggest pain points today? Which use cases are most relevant right now? And what are we most likely getting wrong about how agents will buy and sell? Thank you to everyone who built with us and challenged our thinking. The agentic economy will be shaped by builders pressure-testing these systems in the real world.
250
3 条评论 -
Maia Josebachvili
Stripe • 2万 位关注者
I had fun chatting with Rana el Kaliouby, Ph.D. about how Stripe is building the economic infrastructure for AI on the Pioneers of AI podcast. We went deep on how we are enabling agentic commerce, how we’re helping AI companies grow, and our culture of experimentation with AI at Stripe. And yes, somehow conversations always seem come back to skydiving 🙃
75
4 条评论 -
Eileen O'Mara
2万 位关注者
The annual letter is a moment for me to reflect on the year we’ve had. Businesses on Stripe processed $1.9 trillion in total volume, up 34% from 2024, and equivalent to roughly 1.6% of global GDP. The internet economy is speeding up. Stripe now powers more than 5 million businesses, including all of the top AI companies like OpenAI, Cursor, and Anthropic, many of the largest blue-chip companies (90% of the Dow Jones Industrial Average), most of the biggest tech companies (80% of the Nasdaq 100), and a significant fraction of freshly minted startups. For all these companies, we believe optimizing payments infrastructure is an almost guaranteed way to capture extra revenue and is among the highest-ROI growth activities a business can undertake. Microsoft, for example, evaluates the performance of each of its payment service providers monthly, discovering that Stripe delivered a meaningful increase in authorization rates. After switching to Stripe, London Gatwick's payment acceptance rose by 2.5 percentage points. Read our 2025 annual letter to learn why we are feeling as optimistic about the internet economy as ever: https://lnkd.in/eucf-rAK.
936
12 条评论 -
Reid Hoffman
Manas AI • 280万 位关注者
Nan Ransohoff is the head of Public Goods at Stripe, and now they're focused on tackling a problem for all of humanity: respiratory infections. Most otherwise healthy people get sick two to four times a year. It's annoying. We tolerate it because it feels like part of life. But the aggregate cost, including lost productivity and long-term health effects we haven't characterized well, are mounting. The tools to solve this problem have improved significantly over the past couple of years, including new AI tools that can accelerate research and treatment. This is increasingly a tractable problem, not an inevitable one. The question isn't whether we could make respiratory infections a thing of the past. The question is timeframe, resources, and whether we treat it like the orphan it is: too familiar to feel urgent, too distributed to attract a single champion.
302
56 条评论