Find out what life is like for an Intel research team working on future world-changing technologies and solutions.
Intel is not only about current-day business projects. It’s also about working on projects of the future—constantly looking for technological development, even when it’s not always easy or obvious. What is daily work and life like for our research team? To find out, we talked with Marek Landowski from the Extreme Scale Computing (ESC) team, which runs advanced applied research projects within the Office of the CTO, part of the Data Platforms Group (DPG) at Intel Poland.
Tell us about your business group.
We conduct research to explore new possibilities in technology development—or what’s called pathfinding. As a group, we don’t focus on fast fixes or products that are currently available in market or new solutions that could be implemented right away. Our research focuses on technological advances ahead of what’s generally available, and where these intercept Intel’s product lines to offer higher value to our customers.
What excites you most about working in your business group?
We ask questions to which answers are not yet known or where no obvious solutions are available. Those are open problems, which require scientific rigor and engineering spark to come up with a solution. Sometimes, it is two weeks’ worth of work. Other times it takes us months to come up with a prototype that meets challenging goals and is at the cutting edge of corporate research. That’s very exciting. We have freedom to pursue research avenues and to develop our ideas into new concepts, which are then reviewed to determine if they are worth investing in—all to add value to the disruptive technologies that the ESC team is working on.
What does the whole research and development process look like?
To cut the edge, we are considering solutions to problems broadly, from circuits to applications, so the technology is developed across all the stacks and appropriately optimized. The process can be loosely bucketed into pathfinding, frontend development, backend development, physical design, and workloads.
In pathfinding, we try to understand what the actual problem is, what challenges we need to solve, and come up with ideas—naive or not—about how to solve them. And we assess what is working and what is not.
We begin with analysis of workloads and kernels from customers. This is where we study customer workloads and datasets; consider data sensitivity and understand fundamental elements of algorithms and potential scaling opportunities in relation to the designed architecture.
Frontend development is basically moving pathfinding concepts into early design stages within the RTL environment, getting initial coding and hardware architecture specification matured. It also consists of the black box template functions and black box connectivity, moving down the stage of implementing things in RTL.
The software phase follows a similar approach; identifying black boxes and tools with IP regions and how they are going to connect and communicate on both the RTL and software sides. The development process continues to iterate those and goes deeper. On backend development in RTL, you have validation and floor planning getting ready for synthesis with validation plans which are becoming more sophisticated and extensively targeting functionality and performance. In software, validation becomes a big component with an enormous test plan, whereas development moves from architecting prototype and proof of concept into a working-at-scale software solution.
From there, it is a physical design—boards, sockets, cooling, power delivery, configuration—how we are going to have it in a rack and physically wire it in a limited lab space while at software, it follows with documentation write-up and development of firmware, drivers – and all the fun stuff that nothing will work without once we power on the machine.
What is the most difficult aspect of daily work for you and your team?
Having a constantly open mind and searching for different angles to approach the problem at hand. Usually, influenced by bias or past experience, we realize something is not enough to be disruptive, and difficulty comes with the need to look out of the box at a variety of perspectives and pick the one which has the most chance of cutting the edge. The challenges we face go beyond the limits of our knowledge, requiring us to constantly learn, follow scientific publications, and meet with other experts to broaden ideas.
It sounds like a difficult task, not one for everybody. Who are you looking for to join your team?
Sometimes our work is like searching for a needle in a haystack. You must be rigorous while searching for answers and persistent when experiments fail, when some would give up. Curiosity about the world and an open mindset helps a lot. This work is for people who do not want to sit in one place, are inquisitive, negate the status quo—people with engineering spark.
On the other hand, a methodical approach is also important, because experiments are repeated several times to gain certainty about observed results. The ability to sense the outcome, derive insights, and take actions based on them are the key skills needed for research-driven projects. We are looking for people who have already dealt with research or conducted experiments for their own pet projects. We need people with technical knowledge who can work in a multithreaded work environment. Programming skills are important, mainly in C and C ++, because we work close to the hardware. Also fluency in Python as this is usually a language of customer workloads.
What would you say to people who want to apply to Intel Poland?
It’s good to work for a company that makes its own products and searches its own ways in cutting edge technologies. Intel offers opportunities for development of both technical and soft skills – and is also supportive of ideas for technology incubation. If you like an international ecosystem with different cultures, diversity of thought, ideas, and various technologies, Intel Poland may be an interesting option for your future.