Aquarium is going to IROS 2022 in Kyoto! Drop by booth #48 if you want to talk about robots, deep learning, or aquatic machine learning :)
#iros2022
Aquarium
@aquariumlearn
Aquarium helps teams streamline the process of finding issues, validating fixes, and adding data for their machine learning datasets.
aquariumlearning.comJoined April 2020
Aquarium’s Tweets
We're hiring for an infrastructure engineer at Aquarium! We deal with a lot of data - ingesting it, searching it, crunching it, and improving it. And we're looking for someone who can help us build our core internal infrastructure and systems!
1
We're hiring for a frontend platform engineer at Aquarium! Lot of interesting challenges building data visualization + improvement workflows for machine learning. We're looking for someone who can own the technical architecture for our frontend stack.
1
2
New video walkthrough of Aquarium! Aquarium helps teams streamline the process of finding issues, validating fixes, and adding data for their machine learning datasets.
youtu.be/_zK_z8BhlLg
#machinelearning #deeplearning #dataoperations
1
1
Excited to talk more about our partnership with Zesty, it's been great working together!
If you'd like to try Aquarium, please reach out. Get more improvement for your ML models with less time and labeling cost!
Aquarium is going to be at CVPR this year! Drop by booth #1306 at the Expo if you'd like to say hello~
#CVPR2022
2
We just launched our new website! Excited to help out amazing customers make their ML systems work in the real world. Let us know if you'd like to try Aquarium for yourself!
1
2
Getting you ML model to work is one thing, but what about actually deploying it?
Peter Gao shares lessons he learned the hard way doing deep learning in production:
👇
thegradient.pub/lessons-from-d
9
24
There's a lot of talk about "data-centric ML" these days. Here's some practical tips on how you can set up a process to find and fix problems in your machine learning data!
1
2
Launching Aquarium Segments!
- Track your ML model performance on interesting parts of your datasets
- Use similarity search to find more examples
- Set up regression tests to make sure your model is improving
2
We're looking for a world class Head Of Operations to help our company grow!
jobs.lever.co/aquarium/8771c
1
Excited to collaborate with Comet!
1
1
New demo video! Aquarium users can go through the full data curation flow (finding mistakes, searching through unlabeled datasets to find the best data to label, dispatching data to labeling, and updating their dataset) without needing to write any code.
2
2
What are the current bottlenecks preventing the adoption of autonomous driving? and discuss the technical, logistical, and ethical issues around a long-anticipated technology. buff.ly/3hIZNiL
3
9
Finally get to talk about our partnership with - it's been great working together. If you're interested in trying Aquarium for yourself, reach out to us and we can help you ship better models faster!
2
Data operations is a vital but underrated part of the ML team. Shoutout to some of the great data operations folks I've worked with in my career! Share and tag some data operators that you want to show some appreciation for.
1
1
3
A lot of our customers ask us about how to set up an ML data labeling system, so we wrote a guide! If you're ever wondering whether to build it all in-house or aren't sure how to decide between vendors to go with, check this out!
1
3
🎧 New Data Stack Show just dropped: bit.ly/2PTWP0Z
Tune in to hear Peter Gao [], Co-Founder & CEO of , talk about the foundational importance of #data in #machinelearning.
#ai #podcast #datastackshow
2
3
Our CEO went on the Data Stack Show to talk about machine learning, robots, and Aquarium! Let us know if you'd like to try Aquarium for yourself :)
1
4
If you're looking to improve an ML model, it's not enough to just get more data, you need to think about getting the *right* data. Announcing Collection Campaigns, a feature in Aquarium that helps make targeted data collection a lot faster and easier!
5
We wrote a blog post on neural network embeddings, a cool technology that allows you to browse complicated datasets in a nice visual way and automate common actions in your ML workflows.
3
Aquarium was mentioned in Base Case's latest newsletter! Let us know if you'd like to try Aquarium out for yourself :)
3
6
24
36
Excited to announce our seed fundraising led by Sequoia with participation from Y Combinator! Here's a post from our CEO,
2
14
38
Excited to talk about some of the great work we've done with Sterblue! If you'd like to use Aquarium to improve your own models, let us know :)
When you're working on a machine learning project, it's tempting to throw your hands up and say, "let's just label more data and see what happens." However, it's not about labeling more data, it's about labeling the *right* data to improve your model!
2
4
We just launched our new website for Aquarium!
A machine learning model is only as good as the data it's trained on, and Aquarium has been helping a lot of ML teams improve their model performance by improving their datasets.
1
2
Aquarium allows users to quickly search within massive unlabeled datasets. Find more examples of difficult edge cases, send them to a labeling provider, and retrain your model on them. Get the most model improvement for the least labeling cost!
Aquarium now supports plotting map data so you can understand the geographic distribution of your ML datasets! Let us know if you'd like to try using Aquarium for your own use case :)
As your ML team gets bigger, it becomes harder and harder to organize everyone effectively! Here's some tips on scaling your org so your workplace doesn't feel like Westeros.
Our team has worked with a lot of ML teams through Aquarium and from building ML teams at previous jobs. Here's some of the best practices we've learned over time for building and scaling a super-effective ML team!
1
1
1
Our cofounder and CEO went on Forward Thinking Founders to talk about Aquarium and our mission!
1
We just released our first post on the new Aquarium blog!
Today we talk about how the best way to improve your ML models is... wait for it... by improving the data it's trained on.
4
11






