Featured of the week is Interview With Dr. Borja Peleato, he is from Spain.
What name would you like us to use for the interview?
You can refer to me as Borja Peleato. It is up to you whether you want to put Dr. in front of it or not. Just follow the usual convention for all other interviews. As for my affiliation, I am currently a Visiting Assistant Professor in Electrical and Computer Engineering at Purdue University.
Can you give us a little background about yourself?
I am originally from Pamplona, Spain. I did my undergraduate in Barcelona, at the Universitat Politecnica de Catalunya, and graduated with a double major in Mathematics and Telecommunications Engineering in 2006. I did my thesis on Underwater Communications as a visiting student at the Massachusetts Institute of Technology (MIT). Then I was awarded a “La Caixa” fellowship that allowed me to go to Stanford University to get my Masters and Ph.D. in Electrical Engineering. I graduated in 2013 and my Ph.D. thesis was on signal processing techniques for preserving data integrity challenges in flash memories. During the last year of my doctorate and one year after graduating, I was working in a startup, Proton Digital Systems, which designed flash memory controllers. Finally, I joined Purdue University in January 2014, where I do research in the areas of storage, communications, and optimization.
When I am not working, I like traveling, cycling, and mountain climbing. I also like reading, mostly fiction, but I try to avoid it because once I get hung up on a book, I cannot stop…
How did you get into Electrical/Electronics Engineering and when did you start?
Like all kids, I was lectured on how power sockets were dangerous and I should not play with them. Like all kids, that made me grow some kind of fascination with them until I could not stand it any longer and I stuck a pair of tweezers in the holes. I suffered a shock and got scared at first but, unlike other kids, I tried again a few minutes later to see if the same thing would happen again. Luckily, it didn’t because the fuse had blown… I guess that was my first experience with electricity.
I always knew that I wanted to be a scientist of some kind. I enjoyed challenges, the harder I found a problem, the more I enjoyed working on it. However, when the time came to decide what degree to choose, I felt sad about leaving subjects unexplored. I decided to study engineering because I felt that it kept my options open, and I later chose electrical engineering because it seemed to be the hardest and most mysterious among the engineering branches. Electrical Engineers did all sorts of things that I could not even begin to understand, and I wanted in.
What are you currently working on?
I just recently came back into academia, so I am now building up a research group, finishing up some old papers, and seeking new projects to work on. My current research revolves around flash memory controller design and algorithms, but I am planning to start some projects on communications and mathematical optimization in the near future.
Do you have any noteworthy engineering experiences?
One of my most interesting experiences as an engineer happened before I even become an engineer. I spent the last summer of my undergraduate studies in Peru deploying a solar-powered network of voice and email over HF radios. This network became the only link for several isolated villages with nearby cities.
My team had extensive training from EHAS, the NGO supervising the project, but it was a huge challenge for us nevertheless. We were all students without practical experience and the villages where we were working did not even have electricity, much less internet or phone coverage. We had to make the system work without any outside support, and very limited tools to work with.
It was a very rewarding experience, although the part that the villagers enjoyed most was the light bulb that we installed on the radio. Some of them had never seen one before.
What is the trickiest bug you have fixed?
I don’t remember any particularly interesting one… Most of my work is theoretical or computer-simulated, so I do not have to worry about loose wires or fabrication defects. Some colleagues working in big companies have told me about huge programs that have been patched for decades and are now impossible to debug but, fortunately, I never had to face one of those dinosaurs. I can usually debug my code to narrow down the problem, and then exhaustively analyze the block which is failing to see what is going on. I sometimes spent whole days trying to figure it out, but the cause usually was some corner case that was failing or a stupid mistake, such as a wrong pointer in C.
What is on your bookshelf?
Several course books from the classes that I have taken during my studies, a few Spanish novels (La caverna, La pell Freda,…), some sci-fi novels (1984, We, Ender’s Game,…), and lots of folders with printed notes, research papers, etc. Nowadays it is easier to get information out of the internet than from books, so whenever I need something, I just download and print it.
Can you please tell us a little background about your theses in title with “A signal processing approach to overcome data integrity challenges in flash SSD design”?
Flash memory technology is more efficient, shock resistant, and faster than hard drives. Additionally, its lack of moving parts makes it very compact and easy to scale to large or small capacities. The only obstacle keeping SSDs from replacing hard drives is their cost. In order to make them cheaper, manufacturers are aggressively scaling the technology and packing an increasing number of bits in each memory cell, but the increased density has significantly reduced the reliability of the information stored. The cells interfere with each other, degrade faster, and can lose the information when subject to heat.
In my thesis, I studied some of these failure mechanisms and proposed methods for overcoming them. I proposed using a probabilistic graphical model for tracking the health of the memory throughout its lifetime and adapting the read and write mechanisms to the current health. Specifically, I derived algorithms for adapting the power of the programming pulses, the separation between the voltage levels, and the read thresholds employed for recovering the information. Additionally, I proposed efficient methods for correcting the errors that inevitably happen and studied some of the trade-offs that the architecture of SSDs is facing.
As a Senior Flash Channel Architect what are the unforgettable challenge do you encounter when you become in charge of the signal processing block, LDPC code construction, and flash memory characterization?
The storage channel is similar to many other communication channels, with the only difference that communication happens across time instead of across space. However, the reliability requirements in the storage channel are much more strict than in most other types of communication. Losing a packet in a wireless network is not a big deal, the receiver will just ask for the packet to be re-sent. In memory, however, information loss is completely unacceptable. Missing a few bits can render a whole file unreadable, and it is not possible to go back in time to ask the sender to repeat the write.
My work consisted of characterizing the possible failure mechanisms, however unlikely they were, and designing schemes to overcome them without incurring excessive latency and power cost. Characterizing rare events with a limited amount of resources and designing efficient schemes with extremely low probabilities of failure was definitely challenging. It took me a long time to find the correct way to accelerate the measurements without diverging from a real scenario. Also, the clients could not run a simple test to compare their different alternatives, so we had to educate them to justify why our codes and algorithms would work well in those rare events.
Do you have any hobbies outside of work?
Yes, although I keep changing them so I am not outstanding at any of them. During the last few years, while I lived in California, I used to bike a lot. I had to commute 20 miles each way to work and most days I did it by bike. Now that I have moved to Indiana, where the weather is a bit less bike-friendly, I might have to find something else to do… maybe tennis or rock-climbing. Also, I occasionally enjoy skiing and mountain climbing and, like most people I presume, watching movies and going out whenever I have time.
What direction do you see yourself heading in the next few years?
I see myself working hard in my academic career, climbing the academic ladder. I hope to continue collaborating with the industry, but I wish to stay on this side of the fence. I enjoy teaching and the freedom that I get from choosing my own research projects.
As a professor, what words of encouragement you give to your students?
I have only been supervising Ph.D. students for a couple of months. At this point, my students are excited about starting in a new area of research and do not need much encouragement…
However, I have been teaching for several years, first as a Teaching Assistant and now as faculty, so I have significant experience encouraging students who are struggling with the class. When this happens, I try to make them realize that they like what they are doing, at least most of them. Sometimes they can lose their passion because they see their studies from a “mandatory work” perspective. However, most of them have chosen to study engineering because they enjoy it. Once they take a step back and realize what they are learning and how doing the homework is helping them reach a further goal than just “passing a class”, they usually feel much better.
Is there anything you’d like to say to young people to encourage them to pursue Engineering?
A degree in engineering is like a box of very advanced tools. When you finish, you might not have much experience using most of them, but you do have the basic skills. Then, the world is yours to build things or fix problems. Regardless of what you end up doing or where in the world you do it, you will always have that toolbox with you and a mind configured to pick up one of those forgotten tools, practice with it for a while, and put it to good use.