Activity
5K followers
-
Mingzhen Lin liked thisIt's been a real treat to work on the Universal Commerce Protocol (ucp.dev) with a selection of strong partners in the Shopping tech council in the last few months. Agentic commerce is such a dynamic space right now and I'm excited about how thoughtfully UCP has been built to enable businesses to take advantage of the surging interest in consumer AI. I've lent a few ideas but learned so much more about building protocols through the process. Many thanks to the leadership of Amit Handa and Ilya Grigorik in guiding this process. Checkout Amit's post (and the links there to Ilya's posts) to catch up on the latest UCP news! https://lnkd.in/p/gZ4YtZHhMingzhen Lin liked thisIt's been a big month for the Universal Commerce Protocol (UCP) community 🚀 A new release, a new Payments Tech Council, a new Lodging domain capability, new partners actively building the foundations and more ways to get involved. Quick recap 👇 📦 UCP version 2026-08-25 is live A multi-vertical foundation for retail, food, and lodging, plus grocery and in-store capabilities, 3DS2 authentication and split payments, campaign attribution, and loyalty support. Learn more: https://lnkd.in/gC5kWnd2 🎟️ Want the fun version? Ilya Grigorik created an interactive tour that itemizes all 88 improvements in the release. Such a creative way to explore it, thank you Ilya! 🙌 Link: https://lnkd.in/g4xVWEpq 💳 Payments Tech Council formed Welcome to the inaugural members: Adyen, Ant International, Coinbase, Global Payments, Google, PayPal, Shopify, and Stripe. Up next: new working groups for Tokenization, Buy Now Pay Later, and international markets. Stay tuned for nominations opening soon. In the meantime, more details here 👇 https://lnkd.in/gN3ucJ3v 🏨 UCP Lodging: Booking draft spec Real-time pricing and availability, guest registration, and complex payment schedules for agentic hotel booking. Shaped by the Lodging Tech Council (Amadeus, Booking.com, Expedia, Google, Hilton, Marriott, Trip.com) and now open for public feedback: Link: https://lnkd.in/gqygP4UZ 🤝 New ways to contribute: Domain Working Groups (DWGs) DWGs bring industry practitioners together to solve a well defined, time-bound problem and propose capabilities for respective Domain Tech Council approval. Any group of organizations, sponsored by 1+ Domain Tech Council members can propose one: https://lnkd.in/gsTmu6Pz 🔍 Example: Bulk Product Discovery DWG - provides a standardized way for businesses to publish full catalogs as bulk feeds. Proposed by Meta and Google, with Commerce, Channable, Feedonomics, PayPal, Shopify, Shopline, Shoplazza, Stripe, and Wayfair. Tune in to the discussion: https://lnkd.in/gvJSWMEB Help us shape the open standard for Agentic commerce! Review a draft, join a discussion, or nominate yourself for an upcoming DWG. 💡Tip: You can catch up on all our public meeting notes and roadmap discussions: https://lnkd.in/gSFSDSkN #AgenticCommerce #UCP #OpenSource #Payments #Travel #Shopping🚀 Announcing UCP Release v2026-08-25! · Universal-Commerce-Protocol ucp · Discussion #775🚀 Announcing UCP Release v2026-08-25! · Universal-Commerce-Protocol ucp · Discussion #775
-
Mingzhen Lin reacted on thisMingzhen Lin reacted on thisWe're hiring! Let's build something fun together next summer 🚀 We're tackling complex, real-world problems in finance, and you'll get hands-on experience building intelligent agents for investment banking and asset management workflows. The group is also hiring for DS, DE and SWE interns 🤘
-
Mingzhen Lin reacted on thisMingzhen Lin reacted on thisAfter nearly 7 years at Meta, I was impacted by the layoffs four months ago. I decided to take a career break—to recharge, clear my mind, and enjoy some much-needed time for myself. I’m grateful for the opportunity to step back, reflect, and look at life from a different perspective. Now, I’m ready for the next chapter. I’m currently exploring Senior Software Engineer opportunities. If you know of a role that might be a good fit, I’d love to connect!
-
Mingzhen Lin liked thisMingzhen Lin liked thisExcited to begin my MBA at Harvard Business School today as part of the Class of 2028! I’m incredibly grateful to Lazard and the Healthcare Team for their mentorship over the past five years and their support as I embark on this next chapter.
-
Mingzhen Lin liked thisExcited to share a milestone for our team's user embedding modeling work! Congratulations to John and the team on getting this into a public paper! 📜 I'm very proud that this multi-year research investment is now reaching a broader audience. We started the user embedding journey back in 2021, worked through ML and engineering challenges year after year, and witnessed the shift from CPU to GPU modeling as we scaled the impact. Amazing outcomes come from strong teams: deep dedication, solid execution, genuine enthusiasm, and unwavering belief in the vision! ✊ I've always believed this is one of the areas where we can truly "talk" with deep neural networks beyond natural language prompts, surfacing latent user demographic signals to unlock the next generation of LLM-oriented personalization as a key building block for the future. Looking forward to the feedback and ideas this paper sparks! ✨ #UserModeling #EmbeddingModeling #RecSys26 #MetaMingzhen Lin liked thisExcited to share that our paper, “Mosaic: A Fleet of User Embedding Specialists for Recommendation at Meta,” has been accepted to the Industry Track at ACM RecSys 2026! 🎉 Mosaic rethinks large-scale foundational user representation as a fleet of complementary embedding specialists, with each specialist capturing a different facet of user interests and behavior. By bringing these diverse representations together within a scalable production framework, Mosaic provides richer and more adaptable user understanding that can benefit a broad range of retrieval and ranking systems across Meta. This work was completed during my time at Meta and was truly a team effort. Congratulations to my co-authors Xian Sun, Xiangyang Mou, Yujunrong Ma, Christina You, Michael Jiayuan He, Hrishikesh Paranjape, Aakarsha Agarwal, and Hong Li. A huge thank-you as well to Ryan Shue, Han Jiang, Yan S., Zhuoran Yu and the broader Mosaic team, and all of our partner teams whose contributions made this project and impact across Meta possible. 📕 Paper: https://lnkd.in/gxu-fgda Looking forward to sharing our work at ACM RecSys 2026! #RecSys2026 #RecommendationSystems #AI #MachineLearning #Personalization
-
Mingzhen Lin reacted on thisSaddened to hear about the passing of my favorite professor at Brandeis. I had the privilege of working with him on my master’s final project before graduation. He was incredibly kind, patient, and generous with his time, and he always encouraged me to think deeper and do my best. Looking back, I truly don’t think I would be where I am today without his guidance. Thank you for your kindness, dedication, and the lasting impact you had on so many students, including me. Rest in peace, Professor Hickey.Mingzhen Lin reacted on thisTo my Brandeis University and higher education network: We have lost an incredible educator and advocate with the passing of Professor Tim Hickey. Having taught with him ih the MKTYP, I saw firsthand how he poured his energy into student success and championed academic access. He brought a rare, contagious optimism to our shared work, treating every student with profound respect and ensuring underrepresented scholars had a seat at the table in STEM. It was an honor to be his colleague in that space. My heart goes out to his family and the entire MKTYP and Brandeis community mourning this monumental loss. A Celebration of Life can be found via: https://lnkd.in/gH8X4kjp #BrandeisAlumni #HigherEducation #MKYTP #STEMEducation
-
Mingzhen Lin liked thisAn amazing team and cross company partnerships helped bring this new experience to life. I'm excited to see it expand across our platform!Mingzhen Lin liked thisExcited to see this come to life! Congrats team! Our New Checkout Experience is focused on making the path from discovery to purchase more seamless and giving people confidence to buy 🛍️ ✨ https://lnkd.in/ef2fzNNM
Experience & Education
Other similar profiles
Explore more posts
-
Tan Wang
Over the past year, my team and I built Pinterest's Model Context Protocol (MCP) ecosystem from the ground up: a central registry, a growing fleet of domain-specific MCP servers, and production integrations across our IDEs, internal chat surfaces, and AI agents. MCP is an open standard that gives large language models a unified way to talk to tools and data sources. At Pinterest, we used it as the substrate for AI agents that can safely automate real engineering tasks, from querying Presto data on demand to diagnosing Spark job failures and surfacing institutional knowledge. A few things I'm particularly proud of: * We designed for security from day one: two-layer auth (end-user JWTs + mesh identities), business-group-based access gating, and mandatory human-in-the-loop for sensitive actions. * We made it easy for any team to ship a new MCP server by creating a unified deployment pipeline that handles all the infrastructure, so domain experts just write business logic. * The ecosystem has scaled to 66,000+ invocations per month across 844 monthly active users, saving an estimated 7,000 engineering hours per month. This was a cross-functional effort. Thank you to everyone on the Agent Foundations, Security Engineering, and Traffic Engineering teams who made this possible, and to our engineering sponsors for their support and guidance. Full write-up on the Pinterest Engineering Blog: https://lnkd.in/e-mH26Bw
7 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content