We’re excited to join VeloDB (Powered by Apache Doris) and CocoIndex in San Francisco on May 20 for a practical evening on building data infrastructure for AI agents. Jaz Samantha Ku will be giving a talk on the future of text-to-insight, and why ontology-enforced graph querying matters as teams bring agents closer to production. As AI agents move beyond demos, the data layer has to keep up. That means giving agents the right context, not just more context. We’ll share how zero-ETL graph analytics fits into that picture. 📅 Wednesday, May 20, 5:30–7:30 PM 📍 Trellis Cafe, 981 Mission St, San Francisco 🎟️ Grab your free spot: https://lnkd.in/gh_SXfFq Big thanks to Savannah (Yun) Whitman from #VeloDB for organizing! #Agents #ContextEngineering #GraphAnalytics #AIInfrastructure
PuppyGraph
Software Development
Santa Clara, CA 3,602 followers
Query your relational data as a graph. No ETL.
About us
A graph query engine for all your data. No ETL PuppyGraph enables you to query your data as a graph by directly connecting to your data warehouses and lakes. This eliminates the need to build and maintain time-consuming ETL pipelines needed with a traditional graph database setup. No more waiting for data and failed ETL processes. Deploy To Query In 10 Mins PuppyGraph's revolutionary query engine eliminates onboarding hassles that comes with a graph database, letting you deploy and start querying within just 10 minutes. Replacing an existing database? Effortlessly drop in PuppyGraph as the replacement, seamlessly connecting to third-party tools without the need for data or code migration. Petabyte-Level Scalability PuppyGraph eradicates graph scalability issues by separating computation and storage. Our auto-sharded, distributed computation effortlessly manages vast datasets. Querying directly from your warehouses, lakes, or multiple data sources, PuppyGraph empowers you to graph-query all your data, instead of only the data loaded into a graph database. 10-Hop Queries In Seconds PuppyGraph delivers lightning-fast results, handling complex multi-hop queries like 10-hop neighbors in seconds. Our patent-pending technology efficiently leverages all computing resources, ensuring exceptional performance. Want to take it to the next level? More machines, more performance.
- Website
-
http://puppygraph.com
External link for PuppyGraph
- Industry
- Software Development
- Company size
- 11-50 employees
- Headquarters
- Santa Clara, CA
- Type
- Privately Held
- Founded
- 2023
Locations
-
Primary
Get directions
Santa Clara, CA 95054, US
Employees at PuppyGraph
Updates
-
𝗪𝗲 𝗽𝘂𝗹𝗹𝗲𝗱 𝘁𝗵𝗲 𝗹𝗼𝗴𝘀. 𝗧𝗵𝗲 𝗜𝗰𝗲𝗯𝗲𝗿𝗴 𝗠𝗲𝗲𝘁𝘂𝗽: 𝗥𝗦𝗔 𝗘𝗱𝗶𝘁𝗶𝗼𝗻 𝗿𝗲𝗰𝗮𝗽 𝗶𝘀 𝗹𝗶𝘃𝗲 🎬🧊 During one of the busiest cybersecurity weeks of the year, we stepped away from the RSA crowds and into the AWS SF office for a technical evening on Apache Iceberg and scalable cyber infrastructure. One thing was clear from the room: Iceberg is showing up in more real production use cases. For RSA Edition, the focus was massive threat data, modern data lakes, investigation workflows, and where Iceberg fits into the cyber data stack. Huge shoutout to the speakers who made the night worth the walk from Moscone: 🧊 Leticia Webb (ClickHouse) 🧊 Colin Gibbens & Paul Agbabian (Splunk) 🧊 Peyman Mani (Cogent) 🧊 Austin Groeneveld (Amazon Web Services (AWS)) 🧊 Yiheng An & Chao Lei (Palo Alto Networks) 🧊 Weimo Liu (PuppyGraph) And thank you to our amazing co-hosts and organizers for bringing the community together: Amy Krishnamohan • Nathan Yee (Amazon Web Services (AWS)), Zoe Steinkamp (ClickHouse), Zhenni Wu • Jaz Samantha Ku (PuppyGraph) 🎉 Couldn’t join us in March? Watch the highlights below 👇 And if you want to keep the conversation going, we’re back in the Bay Area on May 21 with Amazon Web Services (AWS) and Databricks 😎 Grab your spot here: https://lnkd.in/gnjrHCka #ApacheIceberg #Cybersecurity #DataLakehouse #DataEngineering #SecurityAnalytics
-
PuppyGraph reposted this
What if Geoffrey Hinton had never written that paper in 1986? If you pulled that single document out of history, the technology landscape today would look unrecognizable. The chatbot you used this morning might simply not exist. The cancer detection models used by radiologists would be less capable. The voice assistant in your car would go silent. Every gradient update in every neural network since 1986 traces a path back to that one paper on backpropagation. I wanted to see what happens when you stop counting direct citations and start tracing the full downstream graph of influence. I analyzed 250 million papers and 1.8 billion citation edges to find the hidden pillars of science. The result was uncomfortable. 78 percent of the research in the subgraph I indexed is downstream of backpropagation within just two hops. This paper is a hidden pillar holding up the entire industry. But proving this is a massive data engineering challenge. A traditional recursive SQL query cannot handle multiple hops at this scale. The intermediate tables explode, and the database breaks before it ever finishes. The query that proves the point is the same query that crashes the database. I used PuppyGraph to get the answer. PuppyGraph reads Parquet files directly and maps foreign keys into a native graph structure in memory. No ETL is required, and no data migration is needed. A query that would crash a standard relational database finished in 2.3 seconds. I built an interactive demo where you can pull any of these foundational papers out of the graph and watch the modern products built on top of them collapse. Demo in the comments. The most important ideas become infrastructure. And we are good, anyway. Hinton ended up writing that paper
-
-
Still figuring out your May 21 evening plans? Make it an 𝗜𝗰𝗲𝗯𝗲𝗿𝗴 𝗻𝗶𝗴𝗵𝘁 🧊 Databricks, Amazon Web Services (AWS), and PuppyGraph are proud co-hosts of the upcoming Bay Area #ApacheIceberg Meetup. 📅 May 21 | ⏰ 5:00–8:00 PM | 📍 CANOPY Menlo Park The event is almost here, and the speaker lineup is starting to take shape. Before we lock it in, the CFP form is open until 𝗠𝗮𝘆 𝟭𝟮, 𝟮𝟬𝟮𝟲. Got an Iceberg production story, technical lesson, architecture decision, or “we learned this the hard way” moment to share? 🎤 Submit your CFP: https://lnkd.in/gfRj6ssE Come for practical Iceberg lessons, free food (think unagi don!), and a room full of builders who can and 𝘄𝗶𝗹𝗹 talk about table formats all evening. You won’t want to miss this. Guaranteed 😎 🎟️ Grab your free spot: https://lnkd.in/gnjrHCka Big shoutout to the organizers making this happen: Scott Haines and Lisa N. Cao (Databricks), Amy Krishnamohan and Nathan Yee (Amazon Web Services (AWS)), and Zhenni Wu, Weimo Liu, and Jaz Samantha Ku (PuppyGraph) 💫
-
-
Cyber attacks don’t 𝘄𝗮𝗶𝘁 while your security data catches up. Tomorrow, Weimo Liu from PuppyGraph and Yingjun Wu from RisingWave are going live with an architecture demo on streaming data, Iceberg, and real-time graph queries for cybersecurity investigation. 📅 𝗠𝗮𝘆 𝟳 | ⏰ 𝟵:𝟬𝟬–𝟭𝟬:𝟬𝟬 𝗔𝗠 We’ll cover: ⚡ Streaming security data into Iceberg with RisingWave 🧊 Querying Iceberg tables as a graph with PuppyGraph 🔎 Tracing permissions, exposures, and blast radius 🤖 Grounding agents with ontology-enforced graph queries Security questions rarely stop at one alert, one asset, or one identity. They’re connected. If this is the kind of problem you’re thinking about, you won’t want to miss this one. 👉 Save your spot for tomorrow: https://lnkd.in/dQ4gC53m #Cybersecurity #GraphAnalytics #StreamingData #ApacheIceberg #AIInfrastructure
-
-
Changing your graph schema shouldn’t mean 𝗿𝗲𝗹𝗼𝗮𝗱𝗶𝗻𝗴 all your data. But with the traditional graph pipeline, that’s often the tradeoff: 📦 Move data into a graph database 🧩 Define the schema up front 🔁 Reload when the model changes That slows teams down before they even get to the fun part: querying relationships. With PuppyGraph, the data stays in tables while the graph schema sits logically on top. Physically, tables. Logically, a graph. 😎 Check out the clip below to see how this makes graph adoption much easier. 👇 #GraphAnalytics #GraphDatabase #DataEngineering #DataArchitecture #ZeroETL
-
Come wrap up spring with the 𝗕𝗮𝘆 𝗔𝗿𝗲𝗮 𝗜𝗰𝗲𝗯𝗲𝗿𝗴 𝗰𝗿𝗲𝘄 🧊🌸 PuppyGraph is proud to co-host the Apache Iceberg Meetup with Amazon Web Services (AWS) and Databricks, and you’ll want to be in the room for this one 👀 📅 𝗠𝗮𝘆 𝟮𝟭 | ⏰ 𝟱:𝟬𝟬–𝟴:𝟬𝟬 𝗣𝗠 |📍 𝗖𝗔𝗡𝗢𝗣𝗬 𝗠𝗲𝗻𝗹𝗼 𝗣𝗮𝗿𝗸 We’re bringing Iceberg builders together for an evening of technical talks, real production stories, and lessons that usually only come from doing the hard parts yourself. Expect sessions on: 🐾 Iceberg in production, not just in theory 🐾 Architecture choices, technical lessons, and edge cases 🐾 What builders are learning as Iceberg matures Plus plenty of time to swap notes, chat with other builders, and geek out with the Iceberg community. And we’re not letting you leave hungry: unagi don, spicy stir fry, and more (desserts included!) 🍱🌶️ If you’re building on Iceberg, curious about Iceberg, or just want to meet the folks deep in the lakehouse weeds, come hang out. 🎟️ 𝗥𝗲𝗴𝗶𝘀𝘁𝗲𝗿 𝗳𝗼𝗿 𝗳𝗿𝗲𝗲: https://lnkd.in/gnjrHCka 🎤 𝗪𝗮𝗻𝘁 𝘁𝗼 𝗽𝗿𝗲𝘀𝗲𝗻𝘁? Submit your CFP here: https://lnkd.in/gfRj6ssE Big thanks to the organizers: Scott Haines and Lisa N. Cao (Databricks), Amy Krishnamohan and Nathan Yee (Amazon Web Services (AWS)), and Zhenni Wu, Weimo Liu, and Jaz Samantha Ku (PuppyGraph) 🙌 🙌
-
-
AI is changing cybersecurity on both sides: 𝘩𝘰𝘸 𝘢𝘵𝘵𝘢𝘤𝘬𝘴 𝘩𝘢𝘱𝘱𝘦𝘯, 𝘢𝘯𝘥 𝘩𝘰𝘸 𝘵𝘦𝘢𝘮𝘴 𝘳𝘦𝘴𝘱𝘰𝘯𝘥. As models become more capable, they’re powering more sophisticated attacks while also enabling stronger defenses. This puts more pressure on the data layer to keep security context fresh, connected, and ready for real-time detection. That’s what Weimo Liu (PuppyGraph) and Yingjun Wu (RisingWave) will explore in our live webinar on cybersecurity + graph analytics on streaming data. 📅 May 7 | ⏰ 9:00–10:00 AM This session looks at how to keep up: 🐾 Real-time data processing with RisingWave 🐾 Subsecond graph queries on fresh data with PuppyGraph 🐾 Faster detection with multi-hop reasoning across live relationships 🐾 An agent-friendly setup with ontology-enforced graph querying We’ll use a cybersecurity demo to show how streaming data and graph queries can work together for faster investigation. 👉 Save your spot: https://lnkd.in/dQ4gC53m #Cybersecurity #GraphAnalytics #StreamingData #AIInfrastructure #DataEngineering
-
PuppyGraph is honored to be featured in the Google Cloud Lakehouse partner ecosystem. For our team, this was a real moment to pause. Our logo is now sitting next to companies we’ve looked up to for years, in a space we’ve cared deeply about from day one. Having PuppyGraph included in 𝘛𝘩𝘦 𝘨𝘰𝘷𝘦𝘳𝘯𝘦𝘥 𝘰𝘱𝘦𝘯 𝘭𝘢𝘬𝘦𝘩𝘰𝘶𝘴𝘦: 𝘈𝘤𝘩𝘪𝘦𝘷𝘦 𝘪𝘯𝘵𝘦𝘳𝘰𝘱𝘦𝘳𝘢𝘣𝘪𝘭𝘪𝘵𝘺 𝘢𝘯𝘥 𝘶𝘯𝘪𝘧𝘪𝘦𝘥 𝘤𝘰𝘯𝘵𝘦𝘹𝘵 with Ahmet Altay, Vinod Ramachandran, and Sumeet Singh made the moment feel even bigger. We started with a simple belief: teams should be able to query their existing data as a graph, without moving it into a separate graph database. Seeing that idea become part of a broader lakehouse ecosystem means a lot. As AI agents create new demands for context, semantics, and interoperability, we’re proud to help bring graph-powered context to lakehouse data. No data movement. No new source of truth. Just connected context where teams already work. Big thanks again to Jobin George, Talat UYARER, and the entire Google Cloud team for the partnership and support that made this launch possible. 🙌
-
-
🎉 PuppyGraph is proud to be the launch partner for Google Cloud Lakehouse! With support for the Iceberg REST Catalog API. Unveiled today at #GoogleCloudNext 2026, GCP’s new Iceberg-native Lakehouse gives you open, managed #Iceberg tables. PuppyGraph turns those tables into an enforced ontology your AI agents can actually trust. What does this integration gives you: 🐾 Multi-hop graph queries directly on Google Cloud Lakehouse data 🐾 An enforced ontology layer for AI agents 🐾 Sub-second performance at petabyte scale 🐾 openCypher + Gremlin on your existing tables (no graphDB required) 🐾 Built for agentic workloads, security, fraud, and more When relationships matter, tables alone aren’t enough. PuppyGraph adds structure on top of Lakehouse tables so agents can access context, validate queries, and recover when they get things wrong. Big thanks to Jobin George, Talat UYARER, and the entire Google Cloud team for the partnership and support that made this launch possible. 🙌 🎥 Teaser video below 📝 Full blog in the comments
