01 Aug Interview with Mohsen Hejrati of Clusterone
Mohsen Hejrati is co-founder and CEO at Clusterone. Prior to founding Clusterone, he was a research engineer at Waymo and senior researcher at Vicarious. At Clusterone, he is focused on large scale machine learning platforms for applications in life sciences and the autonomous vehicles industry. His research interests span computer vision and machine learning. Mohsen has published in IEEE Conference of Computer Vision and Pattern Recognition (CVPR), European Conference on Computer Vision (ECCV) and Neural Information Processing Systems (NIPS) among others. Read his full bio.
Interview with Mohsen Hejrati of Clusterone
Q: What need is Clusterone addressing?
A: AI and machine learning are becoming cornerstone technologies for scientists and engineers, but access to these technologies is still cumbersome. Machine learning – and deep learning in particular – requires lots of computing power, which is challenging to set up and manage. In addition, not all companies have the necessary expertise to apply machine learning and AI. Clusterone helps them solve those complex engineering tasks and takes care of setup, maintenance, and orchestration of the infrastructure behind AI. We enable scientists and engineers to run experiments – even distributed deep learning code – with the click of a button.
A few decades ago, using a computer to do work was really hard, today this has become the easiest thing in the world. We want AI to be the same way. No fiddling around, no complex setup, it should be a tool to get work done, not a time sink.
Q: What are the products and/or services Clusterone offers/develops to address this need? What makes Clusterone unique?
A: We offer the Clusterone deep learning platform – think of it as an operating system for AI – as well as professional engineering and research services through our Applied AI team. Our platform removes the need for infrastructure setup and orchestration. It also allows teams to collaborate on deep learning experiments and easily reproduce previous experiments. Clusterone is an enterprise-grade platform and we work closely with our customers to make sure they get the most out of their AI investments.
The platform is infrastructure-agnostic, meaning it can run on any on-premises cluster or public cloud, as well as integrate into existing workflows. We help teams to use their hardware more efficiently by optimizing cluster utilization and thus saving our customers time and money.
Q: What is your role at Clusterone and what excites you about your work?
A: I am the CEO and co-founder of Clusterone. The AI industry is incredibly fast-moving, and I believe we’re at a crucial turning point where AI moves from labs into mainstream engineering. Being part of this development is very exciting. AI has huge potential for many industries, but precision medicine and biotech stand out in particular. We want to be part of the cure for cancer, aging, and many other pressing problems that we face today. I think that is very exciting.
Q: When thinking about Clusterone and the domain Clusterone is working in, what are some of the recent breakthroughs that are propelling the field forward and how will they impact healthcare?
A: On the hardware side, the biggest breakthrough has been how capable and affordable GPUs and computing power in general has become. Cloud providers make it very simple to access this computing power today, which is a very important step. When it comes to software, I believe the development of free and open-source resources has been very important. Public datasets and libraries such as TensorFlow and PyTorch make it very easy to get started with AI today.
These developments will have great impact on healthcare through innovative new diagnostic and analysis technologies (e.g. automated cancer cell detection using computer vision). For doctors and hospitals, it is becoming much easier to gain insights from medical imaging and other data they collect. That way, they’re able to help patients more directly and more individually.
Q: What are the short-term challenges that Clusterone and its peers are facing?
A: The market for AI is still young, knowledge about AI, machine learning, and even data science is still not widely available. There’s also a severe talent shortage for AI and machine learning experts. We have to work to educate people about AI and train new talent.
Also, Hardware and Software solutions are evolving at a rapid pace which makes adopting and maintaining any solution especially challenging. At Clusterone, we’re working hard to stay ahead of these developments and to simplify this complexity for our customers by providing a concise user experience across all types of infrastructure.
Q: Is there anything else you would like to share with the PMWC audience?
A: We are excited to see a growing community of life science practitioners focusing on adopting machine learning for the right problems. We see ourselves serving them to build the next breakthroughs in cancer research and personalized medicine.
