Thursday, May 23, 2024

Xyna.AI at MLCon Berlin, Xyna.AI at OpenRheinMain in Darmstadt, Launch of private Xyna.AI GPU instance

Xyna Bulletin #06

Dear customers, partners and friends,

our activities in the field of artificial intelligence are picking up speed! In the second half of the year we will be active at two conferences under the Xyna.AI label: in September at the OpenRheinMain (ORM) in Darmstadt and in November at the MLCon in Berlin.

We have been developing our own solution portfolio for AI applications for some time now: based on the Xyna.com cloud platform, our own private Xyna.AI GPU instance is currently being created for this purpose.

Have fun while reading!

Best regards from Mainz,
Philipp

LinkedIn
Xyna GmbH
Philipp Dominitzki

PS: Do you know colleagues who are also interested in Xyna? Then please pass it on! A short answer with contact details to bulletin_at_xyna.com is enough. However, if you are not interested, a short email will suffice and we will remove you from the mailing list.

Xyna.AI at MLCon Berlin

The MLCon (https://mlconference.ai/berlin/) is a leading conference for machine learning technologies and innovations in the field of artificial intelligence. It will take place from November 27th to 30th in Berlin.

MLCon, Conference & Training

This year we are a gold sponsor! Alexander will give a keynote on "AI beyond the mainstream". He will address the value (and power) of open components at a strategic level to avoid becoming dependent on commercial black box offerings - which is relevant for any company that wants to protect and preserve its own innovations and values.

In addition, Alexander and Moritz offer an interactive session entitled "Pair-Programming with AI?!", in which we look behind the scenes and demonstrate how we realized the GenAI extension XyPilot for assisted deep modeling with Xyna Factory under this paradigm.

MLCon Speakers

At MLCon we will be appearing alongside a number of prominent names from the AI industry "establishment" - it will be exciting to see how our positions are received and discussed. In any case, we are looking forward to a few talk-intensive days in Berlin!

If you want to witness the whole thing - it is currently early bird time, tickets can still be purchased quite cheaply. And if you can't be in Berlin but are interested in our activities: feel free to send us an e-mail.

Xyna.AI at OpenRheinMain (ORM) Darmstadt

Even before the MLCon, we will be at the OpenRheinMain (ORM) Conference (https://openrheinmain.org) for Open Source and Emerging Digital Technologies on September 22, 2023.

OpenRheinMain 2023; September 22nd, 2023; Darmstadt University of Applied Sciences, Germany

As in 2022 we are one of the main supporters (Gold sponsor) of the conference, which has been organized very successfully from the Computer Science Department of Darmstadt University of Applied Sciences for several years.

h_da member of eut

Here, too, we will go into our approach of Xyna.AI in a keynote, in an accompanying session and in an interactive workshop: the open and academic algorithms and models for sophisticated solutions and pipelines, with which applications in the field of artificial Realize intelligence securely and without proprietary black boxes.

The ORM agenda will be published at the end of August. Participation is free of charge, the registration required for this is already open. We would be happy to welcome one or the other newsletter recipient there!

Launch of private Xyna.AI GPU instance

AI needs compute power! The current generation of artificial intelligence, with its extremely large model sizes, was only made possible by advances in CPU and GPU performance.

We also need sufficient resources for Xyna.AI to be able to develop our innovative algorithms. Although we concentrate on methods with high calculation efficiency - we are motivated both economically and ecologically - the necessary computing power is still considerable.

At the same time, we (like other companies) do not necessarily want to let our innovations run with the big hyperscalers (information security and confidentiality). Especially with applications of AI in the area of critical infrastructure, it is of existential importance to consider such aspects from the beginning and to handle the corresponding data sensitively.

Computerchip

That's why we started building our own compute instances for AI on the Xyna.com platform. Equipped with a lot of CPU power and GPU acceleration, they offer us the opportunity to run our Xyna.AI pipelines under our own control and sovereignty over processes and data.

Of course, we cannot compete with Google or AWS in this regard. But already, these instances represent the greatest amount of computing power our company has ever owned. And we were never badly equipped, even in the days of data mining and parallel computing.

At the MLCon in Berlin (see article above) we will explain that we are focusing on open algorithms and models that come from the academic field.

Stanford University Berkeley Unversity of Carlifornia

Some of these are on a par with the commercial standard solutions from providers such as OpenAI and in some respects even superior. Xyna.AI thus fits into our product philosophy that Xyna Factory is based on open components and is even available on GitHub under an OpenSource license.

The same applies here: if you would like to know more, please contact us at any time or send us an e-mail.