HPCwire presents our interview with Nidhi Chappell, General Manager of Azure HPC, AI, SAP, and Confidential Computing at Microsoft. As an HPCwire 2023 Person to Watch, Chappell shares her insights on the evolving HPC cloud market and key trends, including sustainability. She also discusses her role and responsibilities at Azure, and remarks on the exciting debut of ChatGPT, which was propelled by a supercomputer that her team deployed.
Congratulations on your selection as a 2023 HPCwire Person to Watch. Can you tell us about your role at Azure, your areas of responsibility, what is most challenging and most fun?
Thank you so much! It is a true honor to be a part of this group of people. My job is to accelerate the development of science and adoption of AI by enabling access to the best infrastructure and services for our customers. The most challenging aspect of my role is also what’s the most fun! Namely, working closely with customers to understand their most critical research and business problems, and then figuring out how we can bring differentiated solutions to the table to help these customers achieve more. We embrace the challenge and use it to fuel the innovation we bring into Azure and the velocity with which we do so.
Summarize the major milestones achieved last year for your division and briefly outline your agenda for 2023.
We made ChatGPT possible! Really 2022 will go down in the AI history books as the year of ChatGPT, which in many ways is an inflection point not just in the capability of large language models but how AI would get integrated more deeply into our daily lives. To support ChatGPT, Bing Search required us to bring our years of experience deploying and managing supercomputers and accelerate that further. The system that OpenAI used for ChatGPT training was co-designed, deployed and then managed by my team.
2023 is shaping up to be an exciting year. We continue to deploy supercomputers at a massive scale for our customers. This requires us to push the envelope on compute, storage, IO and really all aspects of system performance. We also see this year as a moment when generative AI models will start to revolutionize all sorts of customer scenarios. We’re focused on learning from and with all our customers as they embrace generative AI in their workflows.
Give us your sense of how the HPC cloud market is evolving and what’s driving business in the public cloud. How does Azure differentiate itself as an HPC provider?
Benefits from elasticity and flexibility was clearly what jumpstarted certain parts of the HPC community’s usage of public cloud, including Azure, and it will remain a core driver of changes in market share dynamics. We’re now in the second wave of HPC in Cloud adoption that’s even larger than the first. From the customer side, a key driver is the C-level (CIO/CTO/CEO) leadership of almost every organization wants to have a digital transformation strategy for which Cloud is a first class citizen and mission critical workloads like HPC, AI, and SAP are included. From the Azure side, we’ve also matured our platform over time to the point where not only can we run those key workloads but do so with leadership metrics around performance and cost-efficiency over a given period. This is an area where Azure has certainly differentiated itself, and we work very hard both internally and with a broad set of technology partners to make sure that continues to be the case. Then finally from an ecosystem standpoint, commercial HPC software providers have had enough time and experience to figure out how they can optimally support their customers on the public cloud.
The third wave that has begun is organizations figuring out across their portfolio of operations how they can use AI, especially generative AI, to significantly augment existing workflows or create net new ones altogether. Microsoft obviously has a lot of talent, learnings, and capabilities in the AI space, so we’re a logical partner for those looking to inject that kind of innovation into their own organizations.
Large language models are taking the world by storm. With the launch of the Azure OpenAI Service, are you seeing changes in the types of users and workloads you host on Azure?
You’re right that the level of excitement is something new and unprecedented. Every customer and partner we speak to brings it up. With the Azure OpenAI service, we are enabling enterprise customers to integrate the same capabilities of ChatGPT into their own technology stacks. Now is the time for enterprises to really look at how they can better understand and personalize engagements with their customers. Advancements in LLM allow chatbots to better understand the context behind customer questions as well as give better feedback. In 2023, we are seeing a surge in customers across all verticals as they experiment with how to integrate these capabilities into both existing and new workflows.
What technologies and trends are you most closely tracking?
How we track or push the technology landscape is driven by where our customers want to take their workloads, and because those workloads are inherently diverse and complex we are always looking at things from a systems-level perspective. The reality is we cannot look at just one or a few things. We have to always be driving developments across the stack, from the SoC, up through the node, rack, cluster, datacenter, and even Azure region level.
One trend that is exciting to see because it aligns so closely with Microsoft’s own values and priorities is a growing emphasis from customers on sustainability. This is an area where our investments in renewable energy and highly power efficient infrastructure enable Azure customers to meet their own sustainability objectives.
What inspired you to pursue a career in STEM and what advice would you give to young people wishing to follow in your footsteps?
My parents are both teachers and taught science and math so in some regards that was my whole world growing up. I knew I wanted to get into engineering, but it wasn’t until I had a microprocessor class that I really discovered my fascination with hardware. One of our assignments was to develop the circuit board for a calculator and that was a real pivotal moment for me. I still remember that day! Since then, I have just tried to pursue hard problems and solve them by bringing engineering teams together. I love being challenged and truly enjoy when we can solve some incredibly hard problem that seemed daunting in the beginning. In my opinion, perseverance is a key attribute and I always encourage people to not give up.
Outside of the professional sphere, what can you tell us about yourself – unique hobbies, favorite places, etc.? Is there anything about you your colleagues might be surprised to learn?
I am a mother of two boys and my spouse has a similarly challenging job which makes our lives fun and challenging. We both enjoy traveling a lot and that is what we like to pursue whenever we can. My favorite memory is watching the sunrise over Machu Picchu. It was simply breathtaking! It was also incredible to see the engineering that went into making that place. Recently, I have picked up knitting and crocheting as hobbies. They are an incredible stress reliever. I highly recommend those to anyone.
Chappell is one of 12 HPCwire People to Watch for 2023. You can read the interviews with the other honorees at this link.