About
Hands-on AI and engineering leader (AI Ph.D.) with 20+ years of experience driving…
Articles by Alex
-
The Future of Computer Vision at the LDV Vision Summit
The Future of Computer Vision at the LDV Vision Summit
Over the years I’ve given invited talks all over the world, at industry events and at scientific conferences, and when…
2 Comments
Activity
15K followers
-
Alex Jaimes, Ph.D. shared thisStartupeable acaba de publicar mi entrevista, donde hablamos sobre: 👉 Cuándo dejar los modelos de los labs y entrenar un modelo propio 👉 Qué necesita una startup para venderle AI a una empresa como JPMorgan 👉 Comprar lo requerido y construir lo estratégico 👉 Cómo responder cuando un agente se equivoca Links al episodio: 🎧 Spotify: rebrand.ly/SSP296INV 📺 YouTube: rebrand.ly/SYT296INV 🎧 Apple Podcasts: rebrand.ly/SAPP296INVDirector AI de JPMorgan: Cómo Venderle Agentes a Bancos | Alex JaimesDirector AI de JPMorgan: Cómo Venderle Agentes a Bancos | Alex Jaimes
-
Alex Jaimes, Ph.D. shared thisI recently attended the The Alumni Society Leadership Summit. I’ve been to this event several times over the years (including being an invited speaker)- it’s always inspiring and this year was not the exception. In addition to the main sessions (which included a fireside chat with John Leguizamo, which was fun and informative), there were sessions on entrepreneurship, leadership, investing, and generational wealth. One of the panelists highlighted how even though she had gotten an Ivy League education, when she started working she didn’t know what a 401k was (a US employer sponsored retirement account). It struck me how little most people know even about the basics of personal finance, and how important that is. It’s one of the areas where I hope AI has a positive impact for sure.
-
Alex Jaimes, Ph.D. shared thisI had a really fun time discussing AI with Daniel Porras Reyes for the AI Without Borders podcast. Excellent questions, and excellent summary below. The conversation is still lingering in my head- we covered other exciting topics such as self-recursion and continuous learning. Please share and comment!Alex Jaimes, Ph.D. shared this“The biggest technological change in history is that we are now delegating decisions.” That line from Alex Jaimes, Ph.D. stayed with me. Alex is Head of Conversational AI and Search at JPMorgan Chase, and was previously Chief AI Officer at Dataminr. In this episode of AI Without Borders, we talked about what changes when AI stops being just a tool and starts becoming part of the decision-making process. Three ideas I keep thinking about: Agents don’t create leverage if you plug them into broken workflows. If AI writes more code but humans review PRs the same way, the bottleneck just moves. The real work is redesigning the process around what agents can now do. The hard part is knowing what not to delegate. Some decisions can be fully automated. Some should be AI-assisted. Others still need human judgment. The winners will be the teams that know the difference. Buy vs. build comes down to strategic value. Alex’s framework: if the capability touches your IP, customer experience, sensitive data, or something uniquely core to your product, consider building. If you need speed and the capability is not strategic, buy. Links in the comments
-
Alex Jaimes, Ph.D. shared thisA couple of weeks ago I ran the Brooklyn Half Marathon (1:44:45)- it was a special one because that's the race that got me started running 11 years ago (1:47:11). For the first time, my watch's GPS completely failed and by mile 6 it was saying that I had already ran 12 miles. Interestingly, for that first race in 2015, I didn't even have a running watch. I also got sick a couple of weeks before the race (this year), and it was very hot, so I had to accept the constraints and not try to get close to my best time. Continuing on the AI analogies: - Whatever you deploy in AI will have constraints, and they're not always limitations of the model. In fact, in many cases, they're limitations in the overall system that will limit performance. - As you prepare for launch, you have to adapt to changing conditions and modify your training. In some cases, it may even mean stopping and examining other parts of the system into which the AI model(s) will be deployed, and improving those first. - Performance in production always needs to be monitored, but you should ensure you have various meaningful metrics even if there's only one "key" metric. Off to the next race (and model :))!
-
Alex Jaimes, Ph.D. shared thisLast week I joined the University of Miami’s ECE Industrial Advisory Board. As part of my visit, I attended the Rothberg Catalyzer Engineering Pitch Challenge and the Senior Design Expo. It was great meeting faculty members, Dean Vincent Omachonu and fellow advisory board members- lots of interesting discussions on the future of education and work. I was very impressed by the pitches and demos from the students- if you’re an angel or early stage investor, there lots of very interesting opportunities and incredible talent! Universities will continue to play a major role in society. It’s a true honor to be part of this initiative and to bring my humble industry perspective. AI is transforming everything and the changes are happening much faster than most people realize- and it’s not just coding tools. Universities have the unique challenge of having to balance what they’re best at- building foundations- with preparing students for “now” (except “now” means by the time those students graduate). Computer science is being transformed in major ways, but most universities have not figured out how to adapt. My view is that foundations will continue to be important, but it’s critical to make room for a lot more hands on practical experiences, particularly in using and evaluating what AI tools do. Thank you Pinhas Ben-Tzvi for inviting me and Nevenka Dimitrova for making the introduction!
-
Alex Jaimes, Ph.D. shared thisJPMC has ranked No. 1, across all industries (including big tech), on LinkedIn's 50 best employers to grow your career in the U.S. "JPMorgan Chase is going all in on AI — committing ... to the technology that now touches nearly every function of the bank" (over $17B invested in tech per year). One of the things that's most interesting is that when you think about what JPMC does, it's more than just "traditional" banking: think Chase travel (one of the larges travel agencies in the US), links to dining, entertainment, and many other types of services. On financial services alone, there are huge opportunities on what AI, and in particular AI Agents can do. My team is hiring scientists and engineers- you'll have the opportunity to publish at top-tier venues like NeurIPS, ICML, and ACL—and see that research deployed to a user base of over 80 million customers. Link in the comments #hiring #AIAgents #LLM #NLP
-
Alex Jaimes, Ph.D. shared thisI ran the Boston Marathon this week! The energy was incredible, and it felt special. It's the most prestigious marathon in the world (130th edition of the oldest marathon, requiring runners meet strict qualifying times), and I think the hardest of the 4 majors I've ran (Boston, NY, Berlin, Chicago). I finished it in 3:41:46- a little slower than expected, but pretty good regardless. It's the perfect metaphor for deploying AI. The first few miles are downhill, so many runners make the mistake of starting too fast- it looks easy. But the course is full of challenges, and the very end is as or more critical than the beginning. Having a strategy is paramount to completing the race successfully. In deploying AI, it's easy to start by building mocks and believing in the magic. But the road to production is often full of challenges to achieve the right mix of accuracy, latency, and cost. This requires a clear strategy from the start and many checkpoints and experiments. The hardest part is often what I refer to as the 'last mile delivery' (and in safety critical applications, even the 'last inch')- getting the models to have levels of accuracy that surpass human performance because the cost of errors is very high. A huge thank you to everyone who came out and cheered, to my friends and team (Stride for Stride) and for my family for the support. It made all of the difference!
-
Alex Jaimes, Ph.D. shared thisI’m running the Boston Marathon on Monday. When I started running 11 years ago and learned about Boston I thought I would never qualify. So until last year, I didn’t even try. Then I decided I wanted to do things differently- and I did. Within a year, I ran 2 marathons and qualified both times- the first one wasn’t enough (you may run faster than the qualifying time, but if other runners are faster, you won’t get in). In the second one I was faster than the required time by more than 10 minutes- enough to make it in. Here I am, ready for Monday. It’s well known that to achieve different results, things have to be done differently. Yet, it’s quite astonishing how easy it it is for all of us to forget and get stuck in the same cycle. In terms of AI, for a lot of problems, scaling (larger models) is seldom enough- processes have to change, and like in increasing speed for a race, you need to do diversify the training and the tools you combine in production. This year, I’m supporting Stride For Stride- an amazing non-for-profit organization that aims to make running more diverse. Since its founding in 2018 by Tom O'Keefe, it has supported over 500 runners from over 27 countries. Please help me by contributing, any amount (even what you spend on coffee! :)). The org has already raised $116k!
-
Alex Jaimes, Ph.D. shared thisIt was a lot of fun to speak at the MIT AI Summit last week (Imagination in Action). Very interesting talks from fellow speakers on all things related to AI agents (I didn’t know Marc Benioff could make great whale and dolphin sounds!). On my panel, I spoke about how, if progress continues on agentic capabilities, a lot of “commodity” shopping could be fully automated, disrupting advertising with a lot of ripple effects. It’s very early, but the field is moving incredibly quickly- trust, authentication, security will continue to be of paramount importance. That, in addition to a simple, open framework that acts like an index across agent registries for cross-platform discovery and interaction. On the second point, I was invited to participate in a session with a small cross-industry team led by MIT to come up with such framework (and someone made the Men in Black photo). Also got to do Project NANDA: The Internet of AI Agents yoga. Thank you Ramesh Raskar for the kind invitation to both! Amazing panel with Vijay R. Pradyumna Chari Shyam Nagarajan Robert Bench Peeyush Aggarwal
-
Alex Jaimes, Ph.D. liked thisAlex Jaimes, Ph.D. liked this𝐇𝐨𝐰 𝐅𝐢𝐫𝐦 𝐒𝐡𝐨𝐮𝐥𝐝 𝐚 𝐑𝐨𝐛𝐨𝐭 𝐆𝐫𝐚𝐬𝐩 𝐁𝐞? Can it pick up a ripe peach without bruising it? Matthew Beveridge will present his paper on material-aware grasping, a fundamental problem in Physical AI, at CoRL 2026. A visuotactile model estimates the material properties (mass, elasticity, friction) of the object while grasping, and uses these properties to determine the minimum grasp force required to handle the object. The approach has been demonstrated on a wide variety of fruits and vegetables. Material perception while grasping also allows the robot to perform tasks such as ripeness-based sorting of a produce while it is handled. 𝐒𝐐𝐔𝐀𝐒𝐇: A dataset of 270 instances of 52 classes of fruits and vegetables is now publicly available. For each instance, the 3D shape and material properties (elastic modulus, mass, volume, and surface friction) are measured. 𝐅𝐮𝐥𝐥 𝐕𝐢𝐝𝐞𝐨: https://lnkd.in/g7UJa-iR 𝐏𝐫𝐨𝐣𝐞𝐜𝐭 𝐏𝐚𝐠𝐞: https://lnkd.in/gGpgGDGG Shree Nayar, Columbia University, Columbia Engineering #CoRL2026 #PhysicalAI #Robotics #PhysicalAI #EmbodiedAI #AgriculturalRobots #AgriTech #RobotLearning #RobotManipulation #HumanoidRobot #ComputerVision
-
Alex Jaimes, Ph.D. liked thisAlex Jaimes, Ph.D. liked thisI’m super excited to share that I’m stepping into a new role as SVP & GM, Growth & Innovation at OpenTable! I've been a fan of the platform for many years and excited to build alongside Debby Soo, Amy Wei, 🥄 Robin Chiang, and many others. My focus will be on building the next chapter of growth and accelerating the pace of innovation. I’m excited to work with an incredible team as we create new ways to connect more diners with more of the restaurants they love. If you’re in the industry, curious to learn more, or interested in helping shape what comes next - please reach out. Onwards! 🚀🚀 #OpenTable #Growth #Innovation
-
Alex Jaimes, Ph.D. liked thisAlex Jaimes, Ph.D. liked thisI shared our workshop paper, "An Industry Perspective on Effective Steering of Multi-Agent Systems," at AI CHAOS! at HCOMP. * Typical oversight research asks: what evidence do I need to approve, reject, or modify this AI decision? * We're exploring: what does an expert need to know, and be able to do, to *steer* the behavior of the system going forward? * Other researchers found it refreshing to see this kind of research problem explored in a real, deployed system; one question I got: how can human-AI researchers actually change the system, not just design what it should do? * That's a big advantage of how applied AI works at Dataminr: product, engineering, and research drive towards shared outcomes, with human needs, agent capabilities, and system constraints evolving together in one fast loop. Thanks to my co-authors Isabel (Wenjuan) Zhang and Jeffrey Kinsey, and to the organizers (especially Ujwal Gadiraju) for a great workshop.
-
Alex Jaimes, Ph.D. liked thisAlex Jaimes, Ph.D. liked thisEarlier today, I shared some news with our employees: after nearly 18 years at LinkedIn and Microsoft, I'll be leaving at the end of the year. Below is the note I shared with our teams. None of this would have happened without you, the people who use LinkedIn and Office every day. You found jobs, hired great teams and learned new skills. You wrote the plans, built the spreadsheets and sent the emails that run businesses all over the world. Watching that happen at the scale of billions of people has been the best part of my career. I can't wait to see what this team and this community build next.
-
Alex Jaimes, Ph.D. liked thisWhat exciting times to live in!Alex Jaimes, Ph.D. liked thisFor AI leaders, warning that their technology is dangerous could be a win-win. It makes them seem responsible while making their product look incredibly powerful. That's Thore Graepel's response to the "awfully convenient" alarm bells going off about AI safety right now. After my conversation with Roman Yampolskiy, who told me superintelligence will very likely end humanity, Thore was exactly the person I wanted to sit down with. He spent years at Google DeepMind and has been in the room where these systems are built. My episode with Thore is live now. Link in the comments.
-
Alex Jaimes, Ph.D. liked thisAlex Jaimes, Ph.D. liked thisNYC friends: join us tomorrow evening, Wed Sep 30 at 7pm, at Cornell Tech for Field Conditions: Incubating Media Arts in NYC. Salome Asega (NEW INC), Michael Connor (Rhizome), and Matthew Niederhauser (Onassis ONX) will talk about the current media arts ecosystem in New York and how institutions are enabling new work across art, design, and technology. This event is part of the Design Tech Public Lecture Series. Organized by Jose Sanchez @ Design Tech. Open to the public. Registration required: https://lnkd.in/gabHc4a9 Jose Sanchez Matthew Niederhauser Michael Connor Cornell Tech Cornell AAP NYC Cornell University College of Architecture, Art, and Planning
-
Alex Jaimes, Ph.D. liked thisAlex Jaimes, Ph.D. liked thisAs I step down today as Chairman and CEO of Mattel, Inc., I am filled with gratitude. It has been a privilege to lead this remarkable company over the past eight years alongside a global team dedicated to our mission and purpose and the important role we play in the lives of children and families around the world. Together, we transformed Mattel into an IP-driven play and family entertainment company, strengthened our foundation, and expanded our iconic brands across film, television, digital, consumer products, and experiences. Mattel is in excellent hands with Roger Lynch and its passionate team, and I have every confidence the company will continue to thrive. To our employees around the world, our Board of Directors, customers, entertainment partners, shareholders, and the generations of fans who love our brands: thank you. I will always be #MattelProud.
-
Alex Jaimes, Ph.D. liked thisAlex Jaimes, Ph.D. liked thisI’m starting a new role as General Manager of Stanford Men’s Basketball. There are few things I believe in more than Stanford University and the quality and character of its student-athletes. To help lead the program I have cared so much about for so long is a true privilege for which I’m extraordinarily grateful. Thank you to the many, many great people around the world who have supported me on this journey. I dreamed of playing basketball for Stanford as a 12-year-old who wasn’t good enough at anything to think it was possible, but big north stars tend to light the way to improbable destinations, and I somehow made it to Stanford. Now the dream continues. It’s time to get to work 🌲
-
Alex Jaimes, Ph.D. liked thisAlex Jaimes, Ph.D. liked thisAt Fellows Forum 2026, I had the privilege of hosting 3 distinguished keynote speakers and moderating a timely conversation on “Advances in Continuous Learning, Reasoning, and Automated Research” with Ed H. Chi of Google DeepMind, Weizhu Chen of Microsoft AI, and Josh Tobin of Recursive. We framed these advances as three nested loops: reasoning through a problem, learning continuously from experience, and redesigning the system that performs the learning. My key takeaways: 📍Reasoning moves models from next-token prediction toward “next-idea prediction.” Chain-of-thought works because context is not merely memory; it can act as a program that guides the model’s computation. 📍Continuous improvement is powered by the interaction of compute, higher-quality synthetic and human data, distillation, reinforcement learning, and test-time scaling. Product deployment is essential because real usage creates the feedback and data flywheels needed for improvement. 📍Models and their surrounding harnesses, including tools, memory, agents, evaluators and workflows, will co-evolve. Capabilities first developed in the outer system will increasingly be absorbed into the model itself. 📍Automated research will advance fastest where experiments are measurable and feedback loops are short. Josh’s examples from NanoChat AutoResearch and NVIDIA’s SOL-ExecBench suggest that performance engineering may be one of the first domains in which AI research systems materially exceed human experimentation speed. 📍The human role will increasingly shift from solving every problem directly to deciding which problems matter, asking the right questions and ensuring that self-improvement remains aligned with human values. The three loops may use different techniques and operate on different timescales, but they share the same foundation: acting, evaluating the result and using feedback to improve the next attempt. Watch the session: https://lnkd.in/eaZs9HYt #FellowsForum #ArtificialIntelligence #ContinuousLearning #AIReasoning #AutomatedResearch #FellowsFund Alex Ren, Lucas Sheiner, Elaine Jiang, Fellows FundKeynote + Panel: Advances in Continuous Learning, Reasoning, and Automated Research | Day 1Keynote + Panel: Advances in Continuous Learning, Reasoning, and Automated Research | Day 1
Experience
Education
-
Columbia University in the City of New York
-
-
Research in machine learning applied to multimedia information retrieval (Electrical Engineering Department).
• Built a computational framework for detecting near duplicate images and videos, and a machine learning framework for constructing automatic visual detectors from user input.
• Studied issues related to human indexing of visual information, and developed novel conceptual structures for the classification of audio-visual attributes (MPEG-7).
• Initiated several research…Research in machine learning applied to multimedia information retrieval (Electrical Engineering Department).
• Built a computational framework for detecting near duplicate images and videos, and a machine learning framework for constructing automatic visual detectors from user input.
• Studied issues related to human indexing of visual information, and developed novel conceptual structures for the classification of audio-visual attributes (MPEG-7).
• Initiated several research projects (Kodak, Rochester Institute of Technology, SUNY Buffalo), contributed to writing grant proposals, and implemented several research prototypes (~30K lines of Java/C++ code).
Chair of Graduate Student Advisory Council (GSAC), the sole university-wide graduate student government. -
-
-
-
-
-
Volunteer Experience
-
Mentor
Endeavor
- Present 7 years 8 months
Mentoring, advising, and investing in the highest-impact entrepreneurs worldwide (1.5M high quality jobs created, $1.6B raised, $15B revenues), to create long-term, sustained growth.
-
Advisor (member of Colombia's AI Expert Mission)
Presidencia de la República de Colombia
- 10 months
The Artificial Intelligence Expert Mission consisted of a group of experts that, for 9 months, produced concrete recommendations in the short, medium and long term on a prospective roadmap to achieve the implementation of AI Policies specific to Colombia.
Colombia’s AI Expert Mission was one of the first of its kind in the region, and one of the first to focus on developing measures for the development of education and employment policies for the fourth industrial…The Artificial Intelligence Expert Mission consisted of a group of experts that, for 9 months, produced concrete recommendations in the short, medium and long term on a prospective roadmap to achieve the implementation of AI Policies specific to Colombia.
Colombia’s AI Expert Mission was one of the first of its kind in the region, and one of the first to focus on developing measures for the development of education and employment policies for the fourth industrial revolution.
https://inteligenciaartificial.gov.co/en/mission/ -
Borard Member
Digital Divide Data (DDD)
- 3 years 1 month
High-quality data annotation for AI via socially responsible outsourcing using work-study in Cambodia, Laos, and Kenya. 7K+ youth employed, $15M+ in revenue.
-
Volunteer park ranger
Katios National Park in Colombia
Publications
Recommendations received
7 people have recommended Alex
Join now to viewOther similar profiles
Explore more posts
-
Kaoutar El Maghraoui
I was recently quoted in IBM’s Think article examining a major shift in AI: the rise of open-weight systems and what that means for the industry. As highlighted in the piece, when the most capable systems become open-weight, the center of gravity in AI moves from closed models to shared ecosystems. We are entering a world where the competition is no longer just about model IQ (raw capability), but also model EQ—the personality, empathy, warmth, and user experience that build trust and stronger human-AI interactions. 👉 Read the article: https://lnkd.in/eejv6vb6 #AI #OpenSourceAI #AIEthics #EnterpriseAI #IBM
-
Ramkumar Narayanan
As I get deeper into the world of #AI, I am getting more convinced that experience and maturity of thought process is going to be more valuable as the tools themselves get better at inferencing. The evidence is starting to emerge where experienced practioners are integrating and using AI more effectively in their day to day, be it developers, marketers and so on. This brings up a whole different discussion on how you would disseminate lived experience to the next generation of professionals as AI takes on the role of a junior staff member working in the team.
12 Comments -
Stefano Soatto
Monday we open-sourced AI Functions in Strands Labs, today we shared a blog that hints at how they are key to making AI Agents operate as Universal Task Solvers. Tomorrow we will explain these concepts in the CS Colloquium at Princeton, and Friday at Columbia. Join if in the area. https://lnkd.in/gqyu2b5w https://lnkd.in/gvP8pAre
5 Comments -
James Henderson
How do you scale responsible AI practices across a growing community without losing the nuance that makes safety research effective? Leonard Bereska - UvA AI Safety, this is a common friction for those operating in the AI community. Balancing rigorous academic standards with the need for broader collaboration can often feel like a tightrope walk. AI Coalition was built to address exactly this. We bring together AI practitioners, researchers, and organizations to collaborate on responsible AI development. By creating a structured network, we help mitigate AI-related risks and ensure accountability in decision-making processes, allowing you to focus on the research while we handle the collaborative infrastructure. We’d welcome the opportunity to connect and explore how our network might support your work. Join the coalition at https://ai-coalition.net
-
Les Karpas
AI is creating jobs at every layer of the stack, from energy to applications. At the @World Economic Forum, NVIDIA CEO Jensen Huang described AI as a “five-layer cake,” spanning energy, chips and computing infrastructure, cloud data centers, AI models, and the application layer — with each layer driving growth and creating jobs worldwide. Read the highlights.
-
Wes McKinney
From Steve Yegge on Software Survival 3.0, the Spicy Takes version: "In a world of AI-written software, the selection pressure of constrained compute resources will favor tools that save more cognition than they cost to discover and use — crystallized knowledge and substrate-efficient computation become the ultimate survival advantages." https://lnkd.in/e5UvfPqt
4 Comments -
Roberto Hortal
""Agent-native Architectures" discusses how LLMs operating in a loop with tools can achieve complex outcomes far beyond coding. This concept is fundamental to creating highly adaptive products. I find the focus on atomic tools & emergent capabilities particularly insightful! 👇 https://buff.ly/ErqjDEp #ProductManagement #AI 🤖
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content