About
Activity
2K followers
-
Ben Meadowcroft reposted thisBen Meadowcroft reposted thisAI can write user stories. It can't decide which stories are worth writing. That distinction matters more than most product managers realize right now. The tools are genuinely impressive. AI can draft documentation, synthesize research, analyze usage data, and generate communications faster than any PM could. Teams that aren't using it for these tasks are leaving real productivity on the table. But I keep seeing a version of this that worries me. PMs who are using AI to move faster without asking whether they're moving in the right direction. The execution speed goes up. The judgment stays the same. What AI is commoditizing is the craft of producing output: writing, summarizing, structuring. What it can't replace is the judgment required to decide what's worth producing in the first place. Which customer problem actually matters. Which stakeholder needs to be convinced and how. Which tradeoff to make when the data points in two directions. Those are the skills that separate product managers who are managing a backlog from product managers who are shaping a business. The best PMs I know are using AI to get the documentation done faster so they can spend more time in customer conversations. They're using it to clear the analytical work so they can focus on the strategic call. They're treating it as a way to reclaim time for the parts of the job that require a human in the room. The craft of product management isn't disappearing. It's being sorted. AI is raising the floor for everyone and simultaneously raising the ceiling for people who know where to invest the time it frees up. The question worth sitting with: when AI handles more of your execution work, what will you do with the time?
-
Ben Meadowcroft shared thisBen Meadowcroft shared thisMost AI agent memory tutorials start with: "First, spin up a vector DB..." 🙄 For local-first agents? That's overkill. OpenClaw took a different approach: 📁 A single .sqlite file — no Docker, no credentials, no ops 🔍 Hybrid retrieval — vector search + keyword search combined 🛡️ Total data privacy — nothing leaves the machine ⚡ Graceful fallback if vector extensions aren't available The honest take? SQLite is the right tool until it isn't. Once you hit multi-tenancy or thousands of concurrent agents — that's when you graduate to distributed SQL. Same SQL mindset. Different infrastructure. Full breakdown here 👇 🔗 https://ow.ly/zGRl50YnWmf #RAG #AIAgents #SQLite #DistributedSQL #TiDB
-
Ben Meadowcroft reposted thisBen Meadowcroft reposted this🏔️🚀 WE MOVED MOUNTAINS (OF DATA)! 🏔️🚀 Plaid succeeded at migrating our 234 databases from 100 services to TiDB in record setting time with a team of 6 engineers! 🧨 We decided to replace Aurora Mysql wholesale with TiDB at Plaid in 2023 due to: 1. Poor write scalability 2. Reliability shortcomings 3. High KTLO burden 4. Poor developer ux and velocity 5. End of life support for MySQL 5.7 We were told that our timeline was “too ambitious and a bit crazy” and that’s by design! You need to be ambitious and a bit crazy to take on a multi-year critical project like this (h/t Jean-Denis Greze). In hindsight, Ed Huang (Co-founder of PingCAP) said we’ve migrated more services on a tighter timeline than almost any other customer in the world. 🤯 In the last 12 months alone, 6 engineers moved 41 services onto our new SQL platform. Each service requires thousands of careful steps that we automated through dynamic runbooks. Results: ✅ 600k QPS handling Tier 0/1/2 data. ✅ 650TB of data running smoothly in multi-tenant clusters ✅ Zero downtime upgrades (3x so far), scale in/out, and schema changes. ✅ Lower cost and maintenance burden than the managed service we moved off ✅ Achieved 99.997% SLA for availability of TiDB in 2025 for company level SLAs ✅ Resource Controls protect from noisy neighbors and provide guardrails for novel query behavior ✅ One OKR nailed and marked complete 🫳 🎤 We are building a reliable storage layer for the next 10 years of Plaid, and we’re just getting started. 🧱 Huge shoutout to the Storage team for driving this end to end for the last two years. 💪 Specific shout to Andrew Chen, Mingjian Liu, Brian Xie, Seyoung Kim, Catherine Shen, siyuan he, Lauren McCarty, (me), Lohit Verma as part of the core team! Thanks to our leadership over the years for the project support and encouragement: Joy Zheng, Jipeng Han, Pranav Kantawala, Jean-Denis Greze and Will Robinson. Thanks to our partnership with TiDB, powered by PingCAP, Ed Huang, Max Liu, and especially Michael Zhang. Massive thanks to all our partner teams at Plaid! You trusted us and we delivered this project together 😇 there’s way too many to name but y’all made this possible. Thank you to everyone who planned, reviewed, coded, migrated, and sat through late-night windows with us… so that we never again have to tolerate maintenance window team bonding! 🚀🎉 Reply below if you want to hear what we did differently to succeed! Want to work with the fastest-moving storage team in the industry? We’re recruiting! Join the rocket ship and shoulder your part of this incredible journey. Let’s build the future of finance together. We dig prior database experience, but many of us on the team didn't have it 🤯, and instead we prioritize high ownership, operational excellence, creativity and grit. 💳✨ #FinTech #Plaid #TiDB #ChampionshipTeam
-
Ben Meadowcroft reposted thisBen Meadowcroft reposted thisWith Manus in the headlines following Meta's acquisition, it’s a great moment to look at what’s powering agentic AI platforms behind the scenes. This blog breaks down how Manus 1.5 uses #TiDB X to support autonomous agents that can ship full-stack applications end to end — with elastic scale, schema agility, and strong isolation across thousands of app instances running in parallel. It’s a solid look at the infrastructure required to turn agent hype into production-ready systems. 👉 https://ow.ly/7ECQ30sTAat #AI #AgenticAI #Meta #Databases #GenAI #CloudInfrastructure
-
Ben Meadowcroft reposted thisBen Meadowcroft reposted thisEver wonder what happens when a 25TB MySQL database hits the ceiling? ⚙️ Bling ran into a tough scaling challenge with a mission-critical MySQL/MariaDB setup: • 25TB of data • 15 replicas • 2–5B rows per table • Peak loads around 350K QPS Instead of taking on a risky, months-long sharding rewrite, the team chose to migrate to TiDB. A few highlights from the journey: • 25TB imported in ~12 hours using TiDB Lightning • Writes cut over in a planned 5-hour window • 50 TiKV nodes + 15 TiDB instances handling 50K+ concurrent connections A great example that scaling doesn’t have to mean manual sharding or massive rewrites. Full story and lessons learned here: https://ow.ly/K3MG50Y1fGbMySQL to TiDB Migration: How Bling Scaled a 25TB DatabaseMySQL to TiDB Migration: How Bling Scaled a 25TB Database
-
Ben Meadowcroft reposted thisBen Meadowcroft reposted thisBig thanks to everyone who joined our Amsterdam meetup this week with Amazon Web Services (AWS) and Bolt. 🙌 Packed room, great engagement, and plenty of chats over beers and food — exactly the kind of energy that makes this community fun to build with. Here’s what we dug into: 🔹 AWS on Kiro with Jukka Forsgren A closer look at Kiro, the AI IDE using prototype-based development to help teams ship production-ready applications with AI at the core. 🔹 TiDB with Daniël van Eeden How TiDB’s distributed SQL foundation and built-in AI capabilities support large-scale, production workloads without slowing teams down. 🔹 Bolt with Leandro Morgado Real-world insights on how Bolt runs TiDB without worrying about scale, HA, or ops overhead plus lots of practical lessons from their setup. Appreciate everyone who showed up, shared, learned, and helped make the night a solid one for the Amsterdam data infrastructure community. 🍻
-
Ben Meadowcroft reposted thisBen Meadowcroft reposted thisThe new Deloitte Fast 500 list just dropped and PingCAP is #125 after 675% growth! Great to see distributed SQL and real-time, AI-driven architectures recognized at this level. Huge credit goes out to the teams building and the customers pushing us every day. Proud of the engineering momentum behind TiDB and the work happening across our ecosystem to support AI-first, cloud-native systems. #DeloitteFast500 #TiDB #DistributedSQL #OpenSource
-
Ben Meadowcroft reposted thisIf you want a deep dive discussion about TiDB internals, he is the goto guy.
-
Ben Meadowcroft reposted thisBen Meadowcroft reposted thisDistributed SQL is the future of databases! Classic SQL systems like Postgres and MySQL are rock solid, but only until your app actually takes off. Then it starts: ⚡️Queries slow down ⚡️Replicas lag behind ⚡️Sharding enters the chat ⚡️Weekends disappear To deal with scale, most teams glue together a mess of tools: OLTP, OLAP, queues, caches, pipelines, dashboards… It's duct tape engineering, not architecture. That’s why Distributed SQL has become such an interesting concept. It combines: ✅ full SQL + ACID transactions ✅ horizontal scalability ✅ built-in replication and failover ✅ global deployment ✅ no sharding logic in your app Basically, it behaves like one database, but runs like a distributed system. If you're wondering how that actually works, or why it matters for real-world workloads, I just released a video on this. Watch the full video here 👉 https://lnkd.in/exr5vHkT Here, I walk through: - What distributed SQL is (in plain terms) - How it compares to traditional SQL and NoSQL - A hands-on demo using TiDB, powered by PingCAP (MySQL compatible) - And a look under the hood: Raft, architecture, and storage layers Let me know what you think or drop any questions in the comments. And thanks to PingCap who sponsored the video!
-
Ben Meadowcroft liked thisBen Meadowcroft liked thisThe 2026 RSA Conference was a great week for the Mimic team. We hosted excellent conversations, welcomed some new faces, caught up with old friends, and mixed in a few sporting events for good measure. A big year lies ahead and we can't wait to share what's coming. For now, here's what the conference week looked like. #RSA2026 #Mimic #AgenticAI #Ransomware #Cybersecurity #AppSec #WeAreHiring
-
Ben Meadowcroft liked thisBen Meadowcroft liked thisDisaster recovery to the cloud has always been a top line interest in hybrid cloud strategies--with the newly introduced array-based replication support for VMware Live Site Recovery (formerly "SRM") the full reality is there! Lower the RPOs, RTOs and TCO of the full solution. See how you can implement with Everpure array-based replication and AWS Elastic VMware Service with a blog by Pavel Kovar on my team. https://lnkd.in/dW-rxJJE Andy Reedy Dan Kogan Velina Krasteva David StamenVMware Live Site Recovery for Disaster Recovery and Cloud Migration with FlashArrayVMware Live Site Recovery for Disaster Recovery and Cloud Migration with FlashArray
-
Ben Meadowcroft liked thisBen Meadowcroft liked thisAI can write user stories. It can't decide which stories are worth writing. That distinction matters more than most product managers realize right now. The tools are genuinely impressive. AI can draft documentation, synthesize research, analyze usage data, and generate communications faster than any PM could. Teams that aren't using it for these tasks are leaving real productivity on the table. But I keep seeing a version of this that worries me. PMs who are using AI to move faster without asking whether they're moving in the right direction. The execution speed goes up. The judgment stays the same. What AI is commoditizing is the craft of producing output: writing, summarizing, structuring. What it can't replace is the judgment required to decide what's worth producing in the first place. Which customer problem actually matters. Which stakeholder needs to be convinced and how. Which tradeoff to make when the data points in two directions. Those are the skills that separate product managers who are managing a backlog from product managers who are shaping a business. The best PMs I know are using AI to get the documentation done faster so they can spend more time in customer conversations. They're using it to clear the analytical work so they can focus on the strategic call. They're treating it as a way to reclaim time for the parts of the job that require a human in the room. The craft of product management isn't disappearing. It's being sorted. AI is raising the floor for everyone and simultaneously raising the ceiling for people who know where to invest the time it frees up. The question worth sitting with: when AI handles more of your execution work, what will you do with the time?
-
Ben Meadowcroft reacted on thisBen Meadowcroft reacted on thisCongratulations to Nancy Wang, 1Password on winning the 2026 AI Excellence Award — Cybersecurity! 🏆 Nancy Wang is redefining cybersecurity for the agentic AI era. As CTO of 1Password, she pioneered architectures that allow autonomous agents to act without ever accessing credentials—setting a new industry standard for secure AI. See the full winner list: https://lnkd.in/eAY4DZnv #CybersecurityAI #AIExcellence #ThreatIntelligence
-
Ben Meadowcroft reacted on thisRaised our $55M Series B, but we've already built: ✅ Proprietary HTAP Database (yes that’s OLAP + OLTP, in one) ✅ 100,000x’ed distributed system load in 12 months ✅ Our own custom DSL (like terraform but for operations) ✅ First ever Agentic SDLC We aren’t a Chatbot for X. We solve actual hard problems. We cannot be replaced by the next release of GPT/Claude like the avg YC batch. And the best part is, we’re hiring for every role. Come on aboard! 🚢Ben Meadowcroft reacted on thisAfter 3.5 years of building and absolutely RIPPING Celsius… We’re announcing that DOSS has raised a $55M Series B, co-led by Madrona and Premji Invest, with Greyhound Capital, Commerce Ventures, Pathlight Ventures, and Intuit Ventures as new investors, and continued investment from Theory Ventures, General Catalyst, Contrary, and Mintaka VC. Before starting DOSS I worked on numerous hardware products. Honestly my opinion today is pretty much unchanged: it’s a miracle that any get built at all. I saw firsthand how the physical world runs on systems that were designed by people who’d never use them. Teams would adapt their entire business strategy to fit into the shape of off-the-shelf applications and would spend 4-10x more on consulting time than on the software itself. Worst of all - they’d wait YEARS for implementations to complete which ultimately wouldn’t reflect how the business actually functioned by the time it was live. That's the problem we're solving. DOSS is the Operations Cloud built for the Real World. We sit around your general ledger and model how your business actually runs. Procurement, inventory, orders, finance, all connected, all in real time. We're already seeing it work. Teams with individual people saving 20+ hours a week (that’s half of their time!). Companies growing 3x in a year without adding any headcount to operations. With Operations Cloud, teams are putting up Hall of Fame stats and we’re. just. getting. started. HUGE thank you to our customers, our new investors, our long-time advisors, and most of all the DOSS TEAM!!!! We're building something that can stand the test of time (no small feat in enterprise software). Enabling platform shifts like this happen roughly once a generation and as Jeremy Irons says in Margin Call, “Don’t you understand? THIS IS IT!” ONWARD! There’s work to be done.
-
Ben Meadowcroft reacted on thisBen Meadowcroft reacted on thisExcited to share that the Databricks Infrastructure team will be presenting at SREcon Americas this year (https://lnkd.in/gU2JVAtk). We’ll be diving into some of the hardest problems we’re solving at scale: • Intelligent load balancing for global, latency-sensitive workloads • AI-powered workflows to operate large, stateful systems. If you're attending, come find us at the Databricks booth or join our networking event (https://luma.com/qo6weqsd) — would love to connect with fellow infrastructure engineers and leaders. Behind the scenes, these systems power some of the most demanding data and AI workloads in the world. Designing for multi-cloud reliability and performance is fundamentally hard — from traffic routing to failure isolation to intelligent debugging — and we’re excited to share what we’ve learned. We’re also hiring across our infrastructure teams: 📍 San Francisco Bay Area 📍 Seattle/Bellevue Area 📍 Bangalore, India If you're passionate about distributed systems and large-scale infrastructure, DM me — happy to chat. #Databricks #SREcon #Infrastructure #DistributedSystems #Hiring
-
Ben Meadowcroft liked thisBen Meadowcroft liked this"Systems of Record are dead in the AI world" - random 202X blog poster As the builder of a new age System of Record, this might be the latest hot take I have the most to say about. Come join us for a night of war stories and future thinking with the people who built databases, systems of record, and ai-native software infra. 💾 Host Will Robbins: Founder at Robbins Capital, ex-GP at Contrary 💾 Yours Truly: CTO at DOSS, ex-Builder @ Rubrik 💾 John Kutay: Director of Engineering at Rippling, ex-Leader @ Striim 💾 Kyle Nesbit: CEO at Credible, ex-Architect at Google BigQuery 💾 Taylor Savage: Partner at Halyard Collective, ex-Head of Product @ Retool, Product @ Airtable 📍 Doss Office, SF 🗓️ 3/24 🔗 RSVP link in the comments
Experience & Education
-
PingCAP
View Ben’s full experience
See their title, tenure and more.
or
Licenses & Certifications
-
Certified Scrum Product Owner (CSPO)
Scrum Alliance
Issued -
Pragmatic Marketing Certified
Pragmatic Marketing
Issued -
Sun Certified Java Programmer
Sun Microsystems
-
Sun Certified Web Component Developer
Sun Microsystems
Patents
-
Recovering quarantined information from backup locations
Issued US-12393690-B2
-
Quarantining information in backup locations
Issued US-12393689-B2
-
Systems and methods for continuous data protection comprising storage of completed I/O requests intercepted from an I/O stream using touch points
Issued US-12379953-B2
-
In-Place Data Recovery
Issued US-12299309-B2
-
Data Resynchronization Methods and Systems In Continuous Data Protection
Issued US-12182420-B2
-
Application migration for cloud data management and ransomware recovery
Issued US-12019748-B2
-
Data resynchronization methods and systems in continuous data protection
Issued US-11747995-B2
-
Systems and methods for continuous data protection comprising storage of completed I/O requests intercepted from an I/O stream using touch points
Issued US 11,609,775 B2
See patentExample embodiments relate generally to systems and methods for continuous data protection (CDP) and more specifically to an input and output (I/O) filtering framework and log management system to seek a near-zero recovery point objective (RPO).
-
In-place Data Recovery
Issued US 11,561,718
See patentAn in-place data recovery method and system include receiving a user request to restore a virtual machine to a version corresponding to a first point in time, identifying a first snapshot of the virtual machine based on the user request, generating a second snapshot of the virtual machine, identifying a second data block in the second snapshot that includes modified data derived from data content of a first data block in the first snapshot, generating reverse incremental backup data including…
An in-place data recovery method and system include receiving a user request to restore a virtual machine to a version corresponding to a first point in time, identifying a first snapshot of the virtual machine based on the user request, generating a second snapshot of the virtual machine, identifying a second data block in the second snapshot that includes modified data derived from data content of a first data block in the first snapshot, generating reverse incremental backup data including the first data block, and restoring the virtual machine in-place based on the reverse incremental backup data.
-
Systems And Method For Continuous Data Protection And Recovery By Implementing A Set Of Algorithms Based On The Length Of I/O Data Streams
Issued US 11,500,664
See patentExample embodiments relate generally to systems and methods for continuous data protection (CDP) and more specifically to an input and output (I/O) filtering framework and log management system to seek a near-zero recovery point objective (RPO).
-
Incremental replication of data backup copies
Issued US 11,422,838
An example method of managing and storing data includes pulling and saving a first snapshot of a first state of a target machine running on a cloud service and creating a first virtual machine (VM) package associated with the first snapshot. The first VM package sufficient to instantiate a VM emulating the target machine with the first state on an on-premise destination VM platform. The method includes pulling and saving a second snapshot of a second state of the target machine, wherein the…
An example method of managing and storing data includes pulling and saving a first snapshot of a first state of a target machine running on a cloud service and creating a first virtual machine (VM) package associated with the first snapshot. The first VM package sufficient to instantiate a VM emulating the target machine with the first state on an on-premise destination VM platform. The method includes pulling and saving a second snapshot of a second state of the target machine, wherein the second state occurs after the first state. A second VM package is created and associated with the second snapshot, the second VM package sufficient to instantiate the VM emulating the target machine with the second state on the destination VM platform. The second VM package is created by updating the first VM package according to differences between the first snapshot and the second snapshot.
Other inventors -
-
Data resychronization methods and systems in continuous data protection
Issued US 11,249,655
See patentThe present disclosure relates generally to methods and systems for data resynchronization in continuous data protection (CDP) and more specifically to an input and output (I/O) filtering framework and log management system to seek a near-zero recovery point objective (RPO).
-
Data backup and disaster recovery between environments
Issued US 11,074,143
A data management and storage (DMS) cluster of peer DMS nodes manages data backup and disaster recovery using a service level agreement. The backup data can be replicated from a primary compute infrastructure to the secondary compute infrastructure. For example, primary snapshots of virtual machines of the application in the primary compute infrastructure are generated, and provided to the secondary compute infrastructure. If the primary compute infrastructure is lost, the secondary compute…
A data management and storage (DMS) cluster of peer DMS nodes manages data backup and disaster recovery using a service level agreement. The backup data can be replicated from a primary compute infrastructure to the secondary compute infrastructure. For example, primary snapshots of virtual machines of the application in the primary compute infrastructure are generated, and provided to the secondary compute infrastructure. If the primary compute infrastructure is lost, the secondary compute infrastructure is initialized to replicate the primary compute infrastructure. Applications that were running in the primary compute infrastructure can failover to run in the secondary compute infrastructure. The replicated backup data is used to facilitate the migration. During a failover, the primary snapshots are deployed in the secondary compute infrastructure as virtual machines.
Other inventors -
-
Systems and methods for continuous data protection
Issued US 11,061,601
Example embodiments relate generally to systems and methods for continuous data protection (CDP) and more specifically to an input and output (I/O) filtering framework and log management system to seek a near-zero recovery point objective (RPO).
Other inventors -
-
Application migration between environments
Issued US 10,503,612
A data management and storage (DMS) cluster of peer DMS nodes manages migration of an application between a primary compute infrastructure and a secondary compute infrastructure. The secondary compute infrastructure may be a failover environment for the primary compute infrastructure. Primary snapshots of virtual machines of the application in the primary compute infrastructure are generated, and provided to the secondary compute infrastructure. During a failover, the primary snapshots are…
A data management and storage (DMS) cluster of peer DMS nodes manages migration of an application between a primary compute infrastructure and a secondary compute infrastructure. The secondary compute infrastructure may be a failover environment for the primary compute infrastructure. Primary snapshots of virtual machines of the application in the primary compute infrastructure are generated, and provided to the secondary compute infrastructure. During a failover, the primary snapshots are deployed in the secondary compute infrastructure as virtual machines. Secondary snapshots of the virtual machines are generated, where the secondary snapshots are incremental snapshots of the primary snapshots. In failback, the secondary snapshots are provided to the primary compute infrastructure, where they are combined with the primary snapshots into construct a current state of the application, and the application is deployed in the current state by deploying virtual machines on the primary compute infrastructure.
Other inventors -
Courses
-
Advanced Corporate Finance
MBA 621
-
Business Ethics
MBA 604
-
Business Finance
MBA 520
-
Business Plan Competition Management
MBA 590R
-
Business Strategy
MBA 580
-
Consumer Behavior
MBA 655
-
Corporate Financial Reporting
MBA 501
-
Curricular Practical Training
MBA 598R
-
Customer Relationship Management
MBA 619
-
Data Mining for Business Inteligence
ISYS 555
-
Entrepreneurial Strategy
MBA 669
-
Introduction to Global Management
MBA 584
-
Leadership 1
MBA 505
-
Management Seminar
MBA 593R
-
Managerial Accounting 1
MBA 502
-
Marketing Field Study
MBA 651
-
Marketing Management
MBA 550
-
Marketing Research
MBA 650
-
New Product Innovation
MBA 693R
-
Operations Management
MBA 530
-
Pricing & Positioning Tools
MBA 553
-
Sales Management
MBA 654
-
Social Media Marketing
MBA 693R
-
Strategic Decision Making
MBA 685
-
Strategic Human Resource Management
MBA 548
-
Strategic Simulation
MBA 687
-
Strategy Implementation
MBA 681
Recommendations received
4 people have recommended Ben
Join now to viewOther 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