About
Product leader and advocate for developer communities and open knowledge.
Activity
6K followers
-
Megan Risdal shared thisReally enjoyed this convo with Walter Reade and Alex Shaw! We talked about how Alex's work with Kaggle on rigorous evaluation for the $1M KPrize (& lessons learned) inspired Terminal-Bench inspired Harbor.Megan Risdal shared thisIn the first episode of Kaggle Conversations, Megan Risdal, Product Lead at Kaggle, and Walter Reade, Technical Lead for Kaggle Competitions, speak with Alex Shaw, Member of Technical Staff at The Laude Institute, co-creator of Terminal-Bench, and creator of the Harbor Agentic Evaluation Framework (https://lnkd.in/gzzqRSjj), about how the AI Agent industry is moving towards autonomy, why software engineering benchmarks are prone to contamination, and what it took to build a robust evaluation framework for the $1M Konwinski Prize. Episode highlights: • Why autonomous software engineering is shifting into a machine learning problem • What the $1M Konwinski Prize revealed about coding • How the team applied competition experience to prevent contamination • How Harbor gives the open-source community a unified environment for creating and running agent benchmarks Watch a clip from the conversation 👇
-
Megan Risdal shared thisWe're hiring FDEs! Be part of building, shaping, and defining the frontier of AI -- the heart of which is novel, economically valuable evals -- with strategic customers. :) https://lnkd.in/gt9H8yNX
-
Megan Risdal reposted thisMegan Risdal reposted thisSince launching Kaggle Benchmarks, the community has created over 10,000 tasks to measure AI model capabilities — clear benchmarks that pinpoint where models fall short and give labs the signal they need to train better ones. But until now, creating a benchmark meant working exclusively in Kaggle's notebook editor, not the stack you actually build with. Today, we're launching local development for Kaggle Benchmarks. You can now create, validate, push, run and download tasks directly from your local dev environment — VSCode, Cursor, Antigravity, Claude Code and more — using the write-kaggle-benchmarks skill. Here's what that looks like in practice: • Describe a benchmark in natural language: your agent writes the code, validates it locally, and pushes it to Kaggle. • Run it against every SOTA model in one go: pass/fail, latency, cost, and token counts come back in your agent panel. • Tasks you push join the public Kaggle Benchmarks ecosystem, where labs use them as a signal to improve their models. If you can measure a capability, labs will work to improve it. The more people building benchmarks that reflect the real world, the better models get at the things that actually matter. Install the skill and get started 👉 https://lnkd.in/g7A4N6DE Built something with it? Share your task and workflow by July 1 and tag @kaggle for a chance to win Kaggle swag and a social shoutout.
-
Megan Risdal posted thisKaggle competitions for 15 years: community hillclimbs on one organization's problem to discover new SOTA Benchmarks are the future: entire world discovers capability gaps to exploit & hillclimb Email meg@kaggle.com if you want to work together! The future of AI is what you want it to be. Companies, research labs, and individuals - I want Kaggle to be the place that gives you an active voice shaping how frontier AI is built. 🚀
-
Megan Risdal shared thisExcited to meet up with Bay Area Kagglers this weekend! Join us!Megan Risdal shared this🏆 Bay Area Kagglers - let's link up! This Saturday, May 16, we're hosting a casual lunch meetup for the Kaggle community with Megan Risdal (Head of Product of Kaggle) - a few Kaggle GMs have already signed up! It will be a wonderful opportunity to have an open in-person conversation about the future of the platform and also network with seasoned Kagglers! Whether you're a Kaggle Grandmaster or just a beginner, sign up here: https:// luma.com/c5q74f8e More details about the location will be released via the Luma event page. It will be somewhere in South Bay! See you there! 🙌
-
Megan Risdal reposted thisMegan Risdal reposted this🏆 Bay Area Kagglers - let's link up! This Saturday, May 16, we're hosting a casual lunch meetup for the Kaggle community with Megan Risdal (Head of Product of Kaggle) - a few Kaggle GMs have already signed up! It will be a wonderful opportunity to have an open in-person conversation about the future of the platform and also network with seasoned Kagglers! Whether you're a Kaggle Grandmaster or just a beginner, sign up here: https:// luma.com/c5q74f8e More details about the location will be released via the Luma event page. It will be somewhere in South Bay! See you there! 🙌
-
Megan Risdal shared thisAre Kaggle competitions saturating faster? Here's normalized private leaderboard scores from featured competitions since 2021. I think it's too early to say. What do you think? It theoretically should be getting easier to get started with a good baseline faster, so it's tempting to interpret the higher intercept of the 2025+ data as support for this hypothesis. But it's tough to say for certain. Lots of confounding factors. Unique participants in competitions with points & medals are up 41% YoY. More people = signal extracted faster. This also doesn't account for difficulty which most would probably say is increasing over time (we certainly care about giving a home to tackle tough, frontier problems). This is hard to measure. What else would you look into? You can explore this data yourself in Meta Kaggle. https://lnkd.in/eDhTzn5m
-
Megan Risdal shared thisCan you build an agent that balances exploration, resource management, and adversarial expansion? Join Kaggle's latest 1v1 simulation competition: Maze Runner! Kaggle simulation competitions are an exceptionally fun way to get started with using AI tools to build software. After joining the competition, point your agent (from Claude Code, Antigravity, etc.) to AGENTS.md and together you'll walk through making an agent submission in minutes. From there you can watch it start to play in head-to-head games against other competitors' agents on the leaderboard, get feedback on its performance, and iterate. Shoutout to Bovard Doerschuk-Tiberi for the launch! https://lnkd.in/efTHekkt
-
Megan Risdal shared thisKaggle prizes are a proxy for the value the industry places on open, community-driven AI discovery. We are at an all-time high: >$5M USD. Kaggle helps the world discover what works by decentralizing the question of what to measure in the first place: the frontier of AI belongs to the people. Join us! https://lnkd.in/gzPCCkwQ
-
Megan Risdal reacted on thisMegan Risdal reacted on thistl;dr - I'm leaving Google/Deepmind/Kaggle. I’ve spent over 11 years in 5 totally different divisions in Google. I walked into the Cambridge Office on May the 5th 2015, intimidated, starry-eyed and disbelieving. Turning in my badge and laptop today feels a lot like a graduation ceremony. I might have loved my college years, but that doesn’t mean I am not ready to put my diploma (or 10 year Cookie Monster Doodle) on the wall and move on. In my tenure, I’ve had the opportunity to do and participate in really remarkable and bananas things. I remember hosting 19 men from a very formal major customer (13 of whom were wearing blue button up shirts) in a huge stakeholder meeting… following a lunch in the Chicago cafeteria with jello shots. A marching band started playing fight songs while we went through the slides. After a forum designed to connect and inspire mid-career women, I had free time on a flight and thought “wouldn’t it be funny if” and a few weeks later the first Memegender event (https://lnkd.in/gAAZ2MwW) took the company by storm with joyful and delighted representation. There was a flood of exuberant creativity that changed the trajectory of my career and friendships forever. At Verily, I started my tenure by helping to launch a diabetes management application. Ruth, in the earnings call that quarter, mentioned all three of our initial insurance partners by name, and I knew that I had been instrumental in getting that launched. My own father participated and for the first time brought his diabetes under control - gaining years of mobility and health thanks to the coaching, education and insights Onduo created. Not long after, COVID-19 sprung up and I found myself on the front lines of creating overnight processes to get testing to people while preserving their privacy and data. I physically set up the testing tables at a university in Fall of 2020, so students could learn in the safest environment cutting-edge technology offered. My last big project at Google has been to transform the fear, uncertainty and doubt of a public who has gotten the message that “they better AI or else” into skill and understanding. Over 2 million learners participated live in the no-cost “5 Days” Kaggle courses and many more have read the whitepapers, executed the codelabs or listened to the overviews. I'm heading towards a future I'm really excited about. To my fellow Googlers, I leave you all with my great love and affection - and my hope and trust that you are directing all your efforts to the profitable benefit of humanity. #google #deepmind #kaggle #goodbye
-
Megan Risdal reacted on thisMegan Risdal reacted on thisOur new San Francisco HQ adaption ✨ We still have very little furniture but we do now have a tree of life. 🌳
-
Megan Risdal reacted on thisMegan Risdal reacted on thisKicking off the week strong at the Google DeepMind hackathon in Bengaluru!
-
Megan Risdal reacted on thisMegan Risdal reacted on this"Some of our top Grandmasters haven't written a single line of code in over six months." I sat down with Megan Risdal, who leads product at Kaggle, to talk about what's actually changing now that AI agents can enter competitions, and it wasn't the story I expected. A few things that surprised me: → It hasn't made Kaggle harder. It's made it more accessible. Non-experts on her own team are now competing using agentic tools. → The skill is shifting. The best competitors aren't writing code, they're managing "agent armies" and finding where the models break. → Benchmarks are fragile. The incentive to game them is enormous, which is why saturation-proof, game-based evaluation matters more than ever. → Humans aren't out of the loop. We decide which problems are worth solving — and that doesn't go away. The takeaway I keep coming back to: the winners aren't the ones clicking "continue, continue, continue." They're the ones who know where AI falls down, and bring real judgment to it. Full conversation is live on Geeksblabla Community 👇 (link in the comments) #Kaggle #AI #MachineLearning #AIAgents #DataScience
-
Megan Risdal reacted on thisMegan Risdal reacted on thisKaggle has never been more accessible than it is today 😊 with LLM & your agent TEAM . After sharing my recent Kaggle journey, a few people kindly asked me to share some of the experiences, ideas, and lessons I have picked up along the way. My honest answer is simple: join a competition and start competing. https://www.kaggle.com/ One year ago, I was afraid to compete on a global stage. I assumed Kaggle was only for elite researchers, experienced engineers, and people with far more knowledge than me. But Kaggle has never been more accessible than it is today with LLM & your agent TEAM : The Kaggle community is wonderful. People openly share notebooks, ideas, discussions, mistakes, and solutions. At the same time, LLMs can now act as thinking partners—not only helping us write code, but also helping us understand unfamiliar concepts, investigate problems, debug experiments, and challenge our reasoning. You still need intuition, persistence, hard work, good teammates, and sometimes a lot of luck. AI does not remove the difficulty. It simply makes the starting point much more reachable. The best way to learn about Kaggle competitions is not to study everything first and wait until you feel ready. It is to jump into a competition, make your first submission, read discussions, test ideas, fail repeatedly, and gradually understand how the game works. You do not need to begin as an expert. You become better by competing. That is exactly how my own journey started. I simply kept joining competitions, asking questions, learning from the community, working with great teammates, and pushing through many experiments that did not work. If Kaggle interests you, there has never been a better time to start. Do not be afraid of the global stage. Choose a competition, jump in, and learn by competing. You may be much closer than you think.
-
Megan Risdal reacted on thisMegan Risdal reacted on thisI have 2 announcements to make: 1-I am really bad at taking sabbaticals 2-I am joining Docker as Chief Product Officer Now all your misbehaving AI agents are my problem.
-
Megan Risdal reacted on thisMegan Risdal reacted on thisWhat an unbelievable Kaggle journey in 2026 this has been! I’m proud to have reached a worldwide rank of #15 in Kaggle Competitions, earning 4 gold medals (3 prize and 1 title) and 6 silver medals along the way. https://lnkd.in/gVrdYtgh Some of the highlights include: 🥇 1st Place — Vesuvius Challenge 🥉 3rd Place — 2026 NeuroGolf Championship 🏅 4th Place — BirdCLEF+ 2026 🏅 12th Place — Santa 2025 I’m incredibly grateful to my teammates, the Kaggle community, and everyone who shared knowledge, encouragement, and support throughout this journey. There is still so much to learn and improve—and I’m excited for the challenges ahead! #Kaggle #MachineLearning #DataScience #ArtificialIntelligence #Competitions
-
Megan Risdal reacted on thisMegan Risdal reacted on thisI am proud to announce that I have officially become a Kaggle Competition Master after obtaining my first Kaggle Gold Medal! 🥇 This is the result of three years of hard work, investments on GPU compute, late-night coding, and gold near-misses. I couldn't have achieved this without the help and collaboration of my teammates: Egor Trushin, @[deeper], and @[lucienderubempre]. The challenge was to use AI agents to design the smallest neural networks capable of solving ARC-AGI problems. It was a fascinating opportunity to explore agentic workflows in depth—building systems that work in parallel, share knowledge across tasks, and continuously refine and improve their solutions. I’m so happy to finally reach this milestone.
Experience & Education
-
Google DeepMind
View Megan’s full experience
See their title, tenure and more.
or
Licenses & Certifications
-
-
-
Kaggle Python Tutorial on Machine Learning
DataCamp
Credential ID 32fa0819b3ba3df5db96f2b7fe139e8334383769
Volunteer Experience
Languages
-
French
Professional working proficiency
-
English
Native or bilingual proficiency
-
Cantonese
Elementary proficiency
Recommendations received
1 person has recommended Megan
Join now to viewOther similar profiles
Explore more posts
-
Ryan Gormican
I would teach AI in my field the importance of feedback loops. My two sides of agriculture on one end, and technology on the other, connect through agritech in how we foster systems. Systems are meant to evolve while keeping weight under a structural foundation. And that’s a big part of my role at TensoAI. On the developer side, it’s how we iterate on a codebase. Making sure we add new features while keeping a system that can scale and reflect the market demand. On the AI side, it’s how we iterate on a knowledge base. Filling in the gaps, refining multilingual capabilities, and making sure connections can be drawn beyond the immediate facts. Connections are how we ensure farming is smart and sustainable: by planning beyond the immediate harvest. As well, in all this talk of domain-specific knowledge when it comes to AI, we tend to forget our foundational characteristics. Intelligence does grow with us, but we also grow with intelligence. We think fields are separate. When it comes to their core, you will find the ways in which they are the same. That’s why when it comes to agritech, I would continue to teach the importance of feedback loops. Because no matter how deep you iterate, your knowledge always loops back to the beginning. #ArtificialIntelligence #AgriTech #SmartFarming #Agriculture #TensoAI
-
Juana W.
As someone who reads & buys books, I got curious about how data science and machine learning influenced the publishing, marketing and media industry. While looking into this, I found this 2021 blog post by Jiun Kim that mentioned how Audible uses big data and data-driven insights to personalize listening experiences and optimize marketing strategies to further reach and engage with audiences. Looking at the broader landscape, other leading publishers use data in their own ways too. Penguin Random House has built a powerful data platform to bring together sales, marketing, and reader info -- using that to forecast demand and optimize pricing. HarperCollins Publishers leans into predictive analytics to track trends and make marketing smarter. Hachette Book Group focuses on market research and sales data to figure out what’s working for readers. And Macmillan is diving into AI for recommendations and digital pricing. Each has its own style in leveraging data to make publishing & marketing more effective and customer-focused. What’s interesting to think about is how big data and AI might continue to grow in publishing -- in areas outside of recommendation strategies, reader engagement analytics and efficient marketing. https://lnkd.in/eiaraDRE
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More