From our blog

Updates from the team, and insights into the world of data-centric natural language processing.
Announcing our 2.7m€ seed round

"A seed growing into a plant in a comic style."

Announcements
Johannes Hötter| 2023-02-16

Announcing our 2.7m€ seed round

And how we got there (with a failed first product) since our start in 2020.

We are thrilled to announce that we have raised a 2.7m€ seed round co-led by Seedcamp and Faber, with participation from xdeck, another.vc and a number of angel investors.
Bayesian Hyperparameter Optimization

"Waves in an ocean in a cartoon-style."

Knowledge
Divyanshu Katiyar| 2023-02-01

Bayesian Hyperparameter Optimization

Best of both worlds: fast and accurate hyperparameter optimization.

We will cover some of the concepts describing how bayesian optimisation works and how fast it is compared to random search and grid search hyperparameter optimisation methods.​
Active learners with transformers

"A roboter student preparing for an exam in a cartoon-style."

Knowledge
Leo Püttmann| 2023-01-19

Active learners with transformers

See how you can leverage large language models for active learning.

This post will explore how transformer-based machine learning models can be used in an active learning setting, as well as which models are best suited for this task.
How we built bricks

"A wall of colored bricks in a cartoon-style."

Engineering
Leo Püttmann| 2022-12-15

How we built bricks

Our simple backend to power an online playground for modular NLP components.

bricks is an open-source collection of modular and standardized NLP components written in Python. It is designed to fit into any NLP application with ease to bridge the gap between idea and implementation. In this post, we'll show you how we built bricks and how you can use it in your own projects.
Data-centric AI

"A green, digital nucleus in a cartoon-style."

Knowledge
Leo Püttmann| 2022-10-22

Data-centric AI

Why it is here, and why it is here to stay.

When a machine learning model performs poorly, many teams intuitively try to improve the model and the underlying code - let’s say switching from a logistic regression to a neural network. Knowing that this can be helpful, it isn’t the only approach you can take to implement your use case.
How qdrant powers our vector search

"Blueprint of an engine in a cartoon-style on purple paper."

Engineering
Johannes Hötter| 2022-10-03

How qdrant powers our vector search

The open-source engine that powers large scale vector search

Embeddings are a generalization of database technologies. Instead of filtering and searching only on structured data such as spreadsheets, we’re currently experiencing search technologies build on top of embeddings. Effectively, you turn text into a query-able structure that embeds the meaning of the text.
How to deploy NLP models to the cloud

"A blue, digital cloud drawn in cartoon-style"

Engineering
Leo Püttmann| 2022-09-23

How to deploy NLP models to the cloud

Bringing a model into production can be difficult. We show you some steps how to do so.

In this article, we want to show you how you can use our refinery Python SDK to quickly extract data from refinery itself, build a NLP model with it and then deploy it to the cloud with the help of Truss, a free-to-use and open-source tool developed by Baseten.
How to finetune your embeddings

"A wall of colored bricks in a cartoon-style."

Knowledge
Moritz Feuerpfeil| 2022-09-15

How to finetune your embeddings

Improve vectors for similarity search and active learning.

We share our experience with fine-tuning sentence embeddings on a commonly available dataset using similarity learning. We additionally explore how this could benefit the labeling workflow in the Kern AI refinery
Beautiful UIs with Figma and Tailwind

"A beautiful purple graphical user interface in a cartoon style."

Knowledge
Johannes Hötter| 2022-07-25

Beautiful UIs with Figma and Tailwind

In this post, we’re going to share how we used Figma and Tailwind to redesign our open-source tool refinery.

The article entirely focuses on how to build beautiful UIs quickly. You don’t need any prior knowledge to understand this post. After this post, you’ll know: (a) Why Figma and Tailwind are such a great combination to build a beautiful UI, (b) How you can quickly build a consistent design, and (c) that those mockups are worth the time! :-)
We're open-source!

"A glass house in a cartoon-style."

Announcements
Johannes Hötter| 2022-07-18

We're open-source!

Being the data-centric sibling of VS Code.

We have been working tirelessly towards this day for a long time. Finally, we can say that Kern refinery goes open-source, and we celebrate this with our version 1.0!

We generated the blog post images using Stable Diffusion.

Common use cases

  • Email automation
  • GPT-like content
  • Data-centric NLP

Product

  • Platform
  • Architecture
  • How it works
  • Labeling services
  • One API for everything

Docs

  • Changelog
  • refinery
  • bricks
  • gates
  • workflow

Company

  • About
  • Blog
  • Careers
  • Contact
  • Pricing
  • Lean NLP Canvas

Other

  • Imprint
  • Privacy policy
  • Terms of service
  • Security
  • Cookie settings

Subscribe to our newsletter

The latest news, articles, and resources, sent to your inbox weekly.

© 2023 Kern AI GmbH. All rights reserved.