Announcing: Project Ymir

TLDR; Heimdall v1 release notes and a deep dive into the Heimdall ML roadmap.

v1.0 Release Notes

The Heimdall team has been diligently monitoring your feedback. We have introduced various patches to elevate your experience with Heimdall ML.

Authentication system upgrades to strengthen overall system security. Forgot password workflow upgrades.
HTTPS enabled with secure TLS for all requests.
Oracle database integration - allow file uploads as a table from a public Oracle database.
User experience redesign of the model build workflow, to build controls around the build process. Intuitive design to allow users to track progress.
API documentation upgrades to include more ways for users to understand how to use Heimdall. Addition of Scribe & Loom videos.
Addition of code editor to model details page to showcase sample code and test API without leaving the app. Making a stride towards more no-code tool.
Backend upgrades to properly add observability and error logging for better debugging experience.
Core ML algorithm upgrades to reduce processing time for medium size data-sets.
General patches & upgrades.

Thank you for your love and support as we slowly bring Heimdall ML out of beta and into a more mature state. On August 17th, Heimdall ML will officially be 3 months old!

Ymir - The future of Heimdall

As we slowly bring Heimdall ML into a more stable state, it's very important to share with the community what our plans are. The next phase of Heimdall's development will be called Project Ymir.

As you may already know, Heimdall, is a major figure in Norse mythology. Heimdall is gifted with the power of foresight and is the guardian of the kingdom of Asgard. At Heimdall ML, we want to bring the ability of foresight to all while protecting your privacy.

Ymir is a primordial giant in Norse mythology and is linked to the creation of the world. Ymir was born from the dramatic encounter between fire & ice. All the norse gods were born from Ymir's body so therefore he served as the building blocks of the Norse universe.

The goal of Heimdall is to bring advanced machine learning to everyone. Our mission is to democratize the access to machine learning. As we pursue these goals, it's important to be able to build powerful algorithms that can process vastly different types of data.

We have been thinking of ways to strengthen our core algorithm: the obvious idea was to add more capabilities into our existing modeling engine. We are currently a very modular platform so although this approach seems reasonable, it doesn't align with our infrastructure design.

With this is mind, today, we are announcing the Heimdall ML roadmap with four modular product lines. These product lines will function independently but also be directly integrated into the core machine learning engine. We have many incredible ideas on how to strengthen our algorithm and we are going to enhance our offering into four different product buckets. Review our about us page to learn more about our core values.

Reading

We live in a world where text data is everywhere. We believe that our clients also have a vast amount of text data, including but not limited to, product descriptions, product reviews, tweets, comments, blog posts etc. There are a wealth of insights hidden within this text data and we think it's important to give Heimdall the ability interpret this data.

Seeing

With the rise of social media, we spend a lot of time looking at images. Our human brains have the ability to interpret these images to gather some context. We want to give Heimdall a similar ability. We want to help heimdall interpret images so that the core modeling engine is able to process images and convert it to a format that Heimdall can model.

Listening

We spend a lot of time watching television and listening to music. Entertainment is rooted in our ability to hear. We hear a joke and we laugh out loud. We listen to our favorite song and are in a better mood. We want to give Heimdall the ability to convert sounds to an interpretable format. We believe that having the ability to understand sound signals will help our clients build intuitive sound based products.

Writing

Although our focus is not on large language models, we believe there are some aspects of large language models that may be useful when modeling outcomes. We as humans have written to express ourselves and we think it may be important to allow Heimdall to write, if necessary. To be clear, we are not building a large language model service, instead we want to create text embeddings in a tabular fashion.

Why four different products?

We believe that in the future everyone will be interested in generating data sets. We forecast the demand for generating data to skyrocket as the focus is shifting towards building stronger artificial intelligence systems. We want to help provide that data in a quick & easy manner.

Currently our core ML engine is built to help organizations add machine learning to their workflows. We don't think that the core ML engine is something that seasoned data scientists would be interested in. But, we believe that data scientists would interested in using a service which outputs clean data sets based on various types of inputs.

We are planning on building our 4 services to output fully cleaned dataset representation of the data that you pass in. This will forever change that way we approach Data Science problems as they exist today.

Closing Notes

The design work for these product lines are already underway and you should expect to see some beta versions of each of these products towards the end of this year.

We cannot express our excitement as we continue to build this service offering out. We are incredibly passionate about data and will always rely on data-driven approaches to build our offerings.

As always, thank you for your continued love & support. If you have any questions or concerns you can always contact us!

Heimdall – Harness the Power of ML

Heimdall is an automatic machine learning platform that gives you (and your team) a ML backend to build incredible user experiences that are powered by machine learning. It offers quick and efficient model building with even quicker model deployment. Heimdall allows your teams to quickly ideate and build out ML products 80% faster.

Heimdall scales from weekend projects to enterprise needs. Contact us today to be ready to compete in the digital age.