Onehouse is a technology company that provides a cloud-native, fully-managed lakehouse service built on Apache Hudi. The company aims to simplify data infrastructure by combining the ease of use of a data warehouse with the scale of a data lake. Onehouse enables users to quickly set up a data lake, process data efficiently, and maintain ownership of their data in open formats, avoiding vendor lock-in.

Onehouse delivers a universal data lakehouse through a cloud-native managed lakehouse service built on Apache Hudi, which was created by the founding team while they were at Uber. Onehouse makes it possible to blend the ease of use of a warehouse with the scale of a data lake, by offering a seamless experience for engineers to get their data lakes up and running. Onehouse offers the widest interoperability for your data in the market across table formats, multiple compute engines and multiple cloud providers.

We have a stellar team of inspired, seasoned professionals including data, distributed systems, and platform engineers from Uber, LinkedIn, Confluent, and Amazon. Our product team has helped build enterprise data products at major enterprises including Azure Databricks. 


Greylock

Addition

Onehouse

🤖 There would be more and better AI in this world if you could store vector embeddings less expensively and manage them more effectively. 

🏡 Now, with Onehouse, you can! Use a universal data lakehouse from Onehouse as the long-term home for your embeddings. Host only the vectors you need for current vector search use cases in your vector database. 

🤓 Chandramouli Krishnan of Onehouse is here to explain it all. 

https://lnkd.in/gf6JKD-2

#onehouse #dataengineering #nolockin #universaldatalakehouse

🔊 Today only, at Data + AI Summit: Kyle Weller and Dipankar Mazumdar, M.Sc 🥑 Mazumdar are speaking on Apache XTable (Incubating) at 5:10pm. 

👀 Also, come see Onehouse  at Moscone Center in booth 130; we have spoken to more than a thousand happy people so far, and today is the last chance!

https://lnkd.in/gY5NzeCM 

#onehouse #dataengineering #nolockin #datalakehouse #apachehudi #apachextable #opensource

Hudi is different.

Most data lake formats work well in bulk.

But Hudi also cares about record-level operations.

Hudi tables use a primary key to uniquely identify each record.

And the "Record Level Index" (RLI) maps record keys to file groups.

This makes these key-based operations faster:
- writes ➜ incremental merge loads
- reads ➜ "point lookup" queries

The index is stored in HFile format.

HFile is optimized for fast lookups.

That's because it uses a structure similar to a "B+ tree".

Let's look at a write example.

Say you want to "upsert" incoming records into a table.

The record keys are first looked up in the index.

If a key exists, the index tells the engine where to find the corresponding record that needs to be updated.

This reduces the amount of data scanned.

Record-level indices are not common in data lake formats.

Hudi only recently introduced it.

Uber uses it to manage some of their large-scale datasets.

Enjoy your day!

#dataengineering #softwareengineering

🗣️ Today only, at Data + AI Summit: Vinoth Chandar and Kyle Weller are speaking on the Universal Data Lakehouse at 4pm. 

👀 Also, come see Onehouse at Moscone Center in booth 130. The Universal Data Lakehouse is at the center of our discussions there. 

🔊 And Kyle and Dipankar Mazumdar, M.Sc 🥑 Mazumdar will be speaking on Apache XTable (Incubating) tomorrow - Thursday - at 5:10pm. 

https://lnkd.in/gVMsVZJK 

#onehouse #dataengineering #nolockin #universaldatalakehouse

⚡️ We are hoping to spark (ha ha) your interest in Data + AI Summit, hosted by Databricks, with the main event kicking off today. 

🏡  Onehouse is at Moscone Center in San Francisco, booth 130, today through Thursday. We got the booth set up yesterday afternoon. Come see us! 

🗣️ Vinoth Chandar and Kyle Weller will be speaking on the Universal Data Lakehouse on Wednesday at 4pm. 

🔊 And Kyle and Dipankar Mazumdar, M.Sc 🥑 Mazumdar will be speaking on Apache XTable (Incubating) on Thursday at 5:10pm. 

https://lnkd.in/gD3hTN_7

#onehouse #dataengineering #nolockin
#lakehouse #datalakehouse #universaldatalakehouse
#apachehudi #apachextable #opensource

AI:ONEHOUSE

Onehouse

Analysis Not Available