Last week the Kao Data team held our 2022/23 kick-off meeting in London and it was great to see so many colleagues in the same room. We’re a really close-knit team and it’s been tough spending the best part of two years working at distance, so to finally all come together was very rewarding.
What made it all the more beneficial was to see so many new faces in the room. I’ve been at Kao Data for just under three years and I’m actually one of the longest serving employees – testament to how fast we’re now recruiting team members to meet the challenge of being a rapidly growing business.
For me personally, working at Kao Data, and alongside our technical and operations teams, led by Gérard Thibault and Paul Finch has been one of the most rewarding parts of my professional career. When I first started at Kao Data it was very small organisation but I was enthusiastic and trusting about how the company could grow, based on Paul and Gérard’s experience and standing within the global data centre industry.
In addition to this, working at Kao Data has also provided the opportunity to work on the cutting edge of technology, and with customers who are transforming the ways in which we live and work.
Organisations such as EMBL-EBI, NVIDIA and InstaDeep, for example, are today utilising high performance computing (HPC) and AI to support and undertake inspirational, life changing research that will have positive implications for humankind on a global scale. As the person responsible for customer implementations at Kao Data, it’s the unique opportunity to work on deployments of this magnitude that makes my work truly meaningful, and my working relationship with our customers so rewarding.
Last year for example, we secured a new customer contract from InstaDeep, who are a global leader in AI powered decision-making solutions for enterprise. InstaDeep is one of the world’s fastest-moving companies who are pushing the very boundaries of reinforcement learning.
InstaDeep has developed a cutting-edge range of AI solutions to solve complex problems across a range of industries including, MedTech, transport, logistics and renewable energy. In just a few years they’ve gone from being an AI start-up to one of the world’s leading names in AI – having twice been named by CB Insights as one of the 100 most promising AI start-ups in the world.
Our newly published case study (within our Resources section) explains some of the ground-breaking research they’re involved with, and why they chose us as their preferred high performance colocation partner. I’ve have been lucky to have been involved with InstaDeep from the very beginning, working hand-in-hand with Nacef Labidi, their impressive Lead DevOps Engineer and Project Lead, as well as his wider team.
As an Elite-level service provider within NVIDIA’s Partner Network, InstaDeep initially handled their processing requirements in-house and via the cloud. However, as a result of their rapid growth, they soon realised they needed to scale their compute in line with demand, primarily because of power, density, cooling and capacity limitations.
InstaDeep approached Kao Data at the start of the third lockdown to commission a bespoke, high-density environment capable of supporting HPC and hybrid cloud requirements, within extremely tight timescales. They needed a new supercomputer which would complement their existing infrastructure solutions, provide the grunt behind their computational training and modelling, and work seamlessly alongside their hyperscale cloud instances.
In response, and in just a few weeks, we were able to design and build a fully-fledged, GPU-accelerated supercomputer, with few on-site personnel (given restrictions) and amid global supply chain challenges, to ensure minimal impact on their customers. Across each step of this journey, we worked as one with Nacef’s team, showing Instadeep that we had the knowledge and experience to do the job, yet were agile enough to meet their critical timescales.
Other notable success across their implementation was firstly our ability to alleviate some legacy connectivity challenges brought about by previous providers and to be able to offer access to a wide range of private and public networks and connectivity providers, with direct on-ramps into the cloud via Megaport.
Secondly, the basis of our data centre design ensures we can guarantee an ultra efficient SLA-backed PUE of <1.2 which helps ensure Instadeep’s deployment remains energy efficient across the long-term. As power prices rise, keeping PUE low is essential to maximise efficient operations as demand increases – something best highlighted by Nacef himself:
“Our supercomputer offers massive compute capacity for our experimentation pipelines and is where all of our development workloads are hosted. The team at Kao Data has supported us to deliver a highly efficient development platform, which is essential to meet current and future demand.”
Recently, at NVIDIA’s GTC event, InstaDeep was named as one of the first start-ups to harness the UK’s most powerful supercomputer – Cambridge-1, which is also housed at our Harlow campus – to power breakthroughs in digital bioinformatics. Two high performance customers, perfectly deployed in the UK’s finest data centre, and now working collaboratively. From a high-performance colocation perspective, it doesn’t get much better than that for me.
With three new data centres either added to our platform or coming soon, and a whole host of new starters joining us – many of whom I was pleased to meet last Thursday, the future looks incredibly bright for Kao Data.
