GENXT

Last week at the Belloch Campus, the ELIXIR BioHackathon gathered 300+ researchers from supercomputing centres, research institutes, biobanks, and innovative companies across the globe to tackle challenges in life sciences research by fostering collaboration, implementing the Global Alliance for Genomics and Health technical standards, and building tools that promote data sharing and interoperability across bioinformatics infrastructures. This year, Alexander Kanitz and Pavel Nikonorov led the BioHackathon Cloud project, focusing on enhancing data privacy and security through #confidentialcomputing adoption. As GENXT, we showcased the #AttestedTLS protocol, implemented in our confidential computing middleware #confido, with demos of: 1. Confidential LLM Inference – leverage LLMs on sensitive data in a confidential cloud where processed data is always encrypted. 2. GA4GH-SDK/CLI + Attested-TLS + ELIXIR TESK – deploy any container in a federated confidential cloud. Other key projects included the WfExS workflow execution system led by Paula Iborra de Toledo from Barcelona Supercomputing Center, JupyterHub integration with TESK led by Viktória Spišaková, cloudifying V-Pipe viral pathogen analyser led by Gordon J. Köhn and Ivan Topolsky, and research on Nebius Soperator led by Javed Habib from Pacific Analytics PTY LTD. A big thanks to GA4GH-SDK contributors Liisa Lado-Villar from CSC - IT Center for Science, Hirotaka Suetake from The University of Tokyo / Sator Inc., and Vincent Emonet from SIB Swiss Institute of Bioinformatics. Special thanks to Davit Shahnazaryan and Davit Harutyunyan from Amaros AI and Nerses Shahbazyan from BostonGene for their contributions to use cases for GA4GH-SDK and ELIXIR TESK enhanced with Attested-TLS. We also extend our gratitude to the Nebius team for generously sponsoring cloud resources with H100 GPUs for the BioHackathon Cloud project. Confidential Computing is reshaping data security in life sciences, and GENXT is at the forefront, pioneering secure, compliant AI and cloud solutions for sensitive data. Despoina Sousoni, Reggan Thomas, Julia Wilson, Jo Mills, Anne Wilcockson, Ravi Kiran M, Mike Bursell, Dmitry Etin, Ilya Burkov, Lilit Nersisyan, Arsen Arakelyan, Hans Binder, Hrachya Astsatryan, Karén Gyulbudaghyan, Ashot Arzumanyan, Aram Adourian, Ayokunmi Ajetunmobi, PhD

Towards Responsible and Confidential AI. This September, the Global Alliance for Genomics and Health, which brings together over 500 organisations, held its 12th Plenary aimed at advancing open standards for secure, privacy-focused, interoperable cloud environments. Pavel Nikonorov, Ravi Kiran M, and Alexander Kanitz co-chaired a GA4GH Connect session “Towards GA4GH-powered Trusted Research Environments”.  As GENXT, we presented the #AttestedTLS protocol which enables users to verify whether remote services enforce their privacy policy. This was demonstrated through a live demo of the GENXT LLM #privateinference service, implemented with our #confidentialcomputing middleware, called #confido. Ravi from the European Bioinformatics Institute | EMBL-EBI and the University of Bristol presented the “Universal Toolbox for Trust-Enhancing Technologies” – a framework to categorise and operationalise the verification of identity, integrity, and policy enforcement in cloud environments. Joris Vankerschaver from Pacific Analytics PTY LTD discussed integrating #Crypt4GH with GA4GH Data Repository Service (#DRS), showcasing how DRS and Crypt4GH standards work together to secure data at rest and during transmission. Athitheya Gobinathan and Aarav Mehta, both Google Summer of Code contributors, presented their projects: Athitheya showcased #Crypt4GH-middleware integrated with the Kubernetes-based GA4GH Task Execution Service (#TES), while Aarav demonstrated the architecture and implementation of #GA4GH-SDK, highlighting its interaction with TES. We are excited to test the cross-integration of ga4gh-sdk, confido, and crypt4gh-middleware at the ELIXIR #BioHackathon2024 in Barcelona. Stay tuned for our updates or reach out to info@genxt.ai if you are eager to contribute!

The landmark moment for confidential computing! This week, Microsoft Azure announced the #generalavailability of Azure Confidential VMs with #NVIDIA #H100 GPUs. At GENXT, we were fortunate to gain early access to these instances at the beginning of the year through the MS4Startups program, which enabled us to develop our #confidentialinference solutions. Now that this technology is available to all, we’re excited to see the incredible advancements it will bring to #ConfidentialAI and #dataprotection.

🚀 We are excited to announce our new partnership! GENXT has joined forces with Scade.pro to secure private data while using LLMs. Scade.pro's no-code platform seamlessly integrates audio, image, text, and video ML models, streamlining workflows. With GENXT Confidential AI solutions, enterprises can safely leverage LLMs as a service while ensuring that all private data remains encrypted and out of reach from any third parties, including cloud and service providers. Together, we’re committed to making AI safer for businesses by protecting private data throughout the AI process. Discover more about our approach to transforming data security.

GENXT team was excited to have been part of the Wellcome Genome Startup School this year! We were thrilled to share insights on the significant and growing role of Confidential Computing in enabling secure data collaboration in genomics. A big thank you for the insightful sessions and the chance to engage with such inspiring speakers and fellow innovators. The exchange of ideas and expertise was truly enriching. Jo Mills, Anne Wilcockson and Alexandra Canet thank you so much for putting together such a great programme. Location: Wellcome Genome Campus