Over the last few months, I've focused my efforts on recruiting activities which means sifting through hundreds of resumes, searching with...
Over the last few months, I've focused my efforts on recruiting activities which means sifting through hundreds of resumes, searching with variously shifting criteria on LinkedIn, video and phone interview calls, organizing onsite visits, checking references, negotiating and assembling offers. This, all done in the attempt to find a few truly exceptional people to add to our team. I’ve emphasized in previous blog articles that I consider recruiting one of the most important things I do leading SRT. It's a lot of people interaction for a person like me who generally leans towards introversion but also among the most exciting and rewarding of my activities. It's a great feeling when I find highly talented individuals who share our passions and interests, align with our culture and choose to be a part of what we are doing — somewhat like finding treasure.
SRT has developed ECHELON from its very inception to run completely on NVIDIA GPUs. It is one of the, if not the single, most sophisticated implementations of a business-critical engineering simulation running on GPUs today. ECHELON’s results, towering speedups over legacy codes and competitive CPU based codes are at least partly due to our choice of hardware platform and they justify our early decision to go all in on GPU technology. Today the choice seems prescient with NVIDIA riding new highs each month and a thriving industry emerging in machine learning and AI precipitated by GPU technology. However, six years ago when the first lines of ECHELON software were being created things were quite different. We had an uphill battle convincing people that GPUs were capable enough and reliable enough for business operations. One of our competitors even published an article claiming that it was pointless to apply GPUs to a complex engineering application like reservoir simulation. This year, 2018, is a different story completely. A consensus is emerging that GPUs are the future of HPC. This conclusion is supported by the steady advances in GPU hardware performance, the stability and maturity of the NVIDIA CUDA software platform and the lack of credible alternatives for accelerating codes, particularly so now that Intel has formally ended the Phi project. What other choice do HPC practitioners have for order of magnitude speed improvements outside of GPUs? SRT has a decade of experience developing sophisticated high-performance applications for GPU. We are privileged to have some of the world’s top experts on our team. If you believe as we do that GPUs are the future of HPC then SRT is a great place to learn that skill from experts while developing one of the most advanced applications in the field.
Productivity is the measure of effort over a period of time. How much can you get done in one day? one week? one year? If you enjoy developing high-performance technical software and would like to spend most of your professional hours doing so I can almost guarantee that SRT is one of the best options for you. Culturally we value developers’ time as the most precious asset we have in the company. We make every effort to allow our developers the ability to focus and we eschew the idea that you can fracture, slice, and dice limited work hours with multiple tasks so that attention is pulled in every which direction. As a small company with a flat management style, there is also less reporting to do, fewer last-minute demands for slides explaining what you are doing to a remote level of bureaucracy. Management knows because his office is right next to you and he’s always available for coffee in the morning. We don’t have disruptive re-orgs that stall productivity every two years because our org is quite simple. In short, much more of your time can be spent doing the things that you enjoy and are productive. We are more flexible in terms of working hours and working location than a large organization can be. Most of the ECHELON development team lives within 10 minutes of the office. So close, in fact, that they frequently enjoy lunch at home with their families. Growing up Italian-American in New York, the importance of family was imprinted on my DNA. SRT provides a way for employees that live far from their extended families to work in ‘deep remote’ locales, meaning that they may work from anywhere around the world for a time while they are visiting their families. There are enormous freedoms one has as a software developer of any kind and we labor to offer those freedoms to employees to improve their quality of life, their work satisfaction, and their productivity. If you are an established individual working on simulator development at a large organization, you are in the prime of your career, you have put an enormous investment into your expertise through training, diligent work, late nights of study, code development and debugging. Ahead of you are potentially the most productive years of your life. To whom will you give those years? Where will you get the most return? This last question leads directly to the final point.
There are actually three types of impact to consider. First, there is the impact of your work on the company and your direct group. In 1992 I was 30 years old and fresh out of my Ph.D. with one child; It was time to get a real job. I was fortunate to land at Exxon where I stayed for about a decade. While I was there I had the opportunity to work with brilliant people on a variety of interesting and challenging problems. However, after a time I found that there was something missing. It became clear to me that it was a long road to the point where I would have a say in the allocation of resources, where I could see my ideas given substance, or I could have a role in hiring. Things moved very slowly. No fault of Exxon but by sheer size, it’s very difficult to make an impact in a company with revenue measured in the hundreds of billions of dollars. This will be true in most large organizations. SRT is about 10 people now and on the way to perhaps 15 by next year. It’s a small intense group of experts who depend on each other for overall success and are focused on a single goal: Create the fastest, most advanced reservoir simulation software in the world. It’s a place where your work and your ideas matter a great deal. When you are one person in a company of 10 your contribution is extremely important. It has impact. Second, there is the group’s impact on the market and the technical landscape. One could have a large impact on the work of a group where the collective effort has minimal long-term effect. I've built my career at the intersection of physics and computing and after almost 30 years I still find it fascinating, challenging and enormously rewarding. The best one can do for professional satisfaction is to contribute to the field they love. At SRT we are developing a unique technology in ECHELON software that is garnering strong interest from a major industry. There is a sense that we are pioneering a new direction, not just in reservoir simulation but in HPC in general. We are early participants in a new epoch of HPC as industry transitions to GPU technology and in doing so we are making a lasting contribution to the field in which we’ve invested our professional careers. Finally, there is the impact you have on your overall well-being and that of your family. At SRT we structure compensation with short, medium and long-term vision. Short term is covered by above market salaries, medium term is covered by year-end bonuses and long-term is covered by ownership through incentive stock options. The simple principle is that as the company prospers so too should the people responsible for its success. There is a strong and direct connection between your hard work and the tangible realization of value that can be accrued personally.
To summarize I end with questions that emerge from my three topics:
Whether you are an experienced individual or fresh graduate, if these messages resonate with you, please get in touch with us and let's talk!
Vincent Natoli is the president and founder of Stone Ridge Technology. He is a computational physicist with 30 years experience in the field of high-performance computing. He holds Bachelors and Masters degrees from MIT, a PhD in Physics from the University of Illinois Urbana-Champaign and a Masters in Technology Management from the Wharton School at the University of Pennsylvania.
What we are doing to help improve the reservoir simulation industry.
Leveraging modern GPUs and neural computing libraries, scientists are using deep learning technology to solve differential equations that emerge from a diverse set of physics-based problems. This blog introduces the topic and explores potential applications and limitations of the technology.Read article →