Interview: Anthony Bak, Ayasdi on How to Get Started on Topology
We discuss the best resources to learn Topology, career motivation, important qualities sought in data scientists and more.
Anthony Bak is a principal data scientist at Ayasdi, where he designs machine learning and analytic solutions to solve problems for Ayasdi customers. Prior to Ayasdi he was a postdoc with Ayasdi cofounder Gunnar Carlsson in the Stanford University Mathematics Department. He's held academic positions at the MaxPlanck Institute for Mathematics, Mount Holyoke College and the American Institute of Mathematics.
His PhD is from the University of Pennsylvania on the connections between algebraic geometry and string theory. Along the way he cofounded a data analytics company working on political campaigns, worked on quantum circuitry research, and studied chaotic phenomena in sand boxes. His friends say that his best idea was to found a College funded cooking club in order to eat food he couldn't afford otherwise.
First part of interview
Second part of interview
Here is there and last part of my interview with him:
Anmol Rajpurohit: Q8. What resources would you recommend to learn and practice TDA? Is Topology adequately covered in Data Science related educational curriculum across universities?
Anthony Bak: TDA is not currently part of any standard data science curricula, but there are a growing number of schools that offer graduate level courses in TDA. Penn, Duke, Stanford, University of Chicago, University of Ohio all come to mind, but there are doubtless others. The perspectives and techniques are not exactly what we do at Ayasdi, but there are many shared ideas. A nice example syllabus with resources: http://web.cse.ohiostate.edu/~tamaldey/course/CTDA/CTDA.html
For self study I also recommend starting with "Topology and Data" by Gunnar Carlsson (http://www.ams.org/journals/bull/
20094602/S027309790901249X/S027309790901249X.pdf). If you're interested in applications to Machine Learning or particular industrial problems, I recommend contacting Ayasdi and asking for the relevant white papers as there is relatively little published in the academic literature. For a more gentle introduction for a working data scientist I have a number of video lectures up on YouTube that I've given at various meetups, academic seminars and industry conferences (sorry for the self promotion).
AR: Q9. What motivated you to work in Data Science?
AB: As a researcher I worked on connections between geometry and string theory, and while I found string theory satisfying in an abstract sense I wanted to work on projects that had more immediate real world impact. I started looking outside of academia, and data science was a draw for a variety of reasons.
I was particularly drawn to the vibrant culture of creativity in data science; it pulls from whatever discipline it needs to without dogma. Data scientists themselves are constantly seeking novel ways to quantify and explain the world around them, and while some people argue about the term "Data Scientist" and say they are really statisticians, analysts, or machine learners, I think that culturally we are distinct and embody the scientific approach to the world in ways that others do not.
AR: Q10. What is the best advice you have got in your career?
AB: "Work it out for yourself. Start with a simple example."
AR: Q11. What key qualities do you look for when interviewing for Data Science related positions on your team?
AB: We look for people with some quantitative background in mathematics or the natural sciences. At Ayasdi, Data Scientists work with customers to solve their problems, do product development, and more fundamental R&D. This is an extremely broad set of responsibilities and we generally expect people to be solid contributors in two of the three areas.
R&D is more important for us than for many data scientists embedded in industry. We work at the cutting edge of TDA and machine learning and can't necessarily rely on academic literature to find solutions for our problems. Some of what we do at Ayasdi challenges machine learning and statistical orthodoxy, and we need people who can independently and creatively think through our methodology and distinguish a heuristic from a theorem.
One thing that we require is that all Data Scientists, even those with a more research oriented bent, work with customers. This keeps all of our efforts focused on solving problems as they appear in the real world and not in idealized settings where something can be proved. As a result, communication with both technical and nontechnical counterparts at other companies is a requirement for a Data Science position at Ayasdi. We ask all interviewees to give a seminar type presentation on a data problem that they've work on. The presentation is evaluated for both content and clarity.
PM me if you have a good way to measure this in an interview.
AR: Q12. On a personal note, we are curious to know what keeps you busy when you are away from work?
AB: I like to explore the Bay Area with my family. While job opportunities in Tech is what brought me here, the cultural and culinary diversity set in stunning physical beauty is what keeps me here.
Related:
His PhD is from the University of Pennsylvania on the connections between algebraic geometry and string theory. Along the way he cofounded a data analytics company working on political campaigns, worked on quantum circuitry research, and studied chaotic phenomena in sand boxes. His friends say that his best idea was to found a College funded cooking club in order to eat food he couldn't afford otherwise.
First part of interview
Second part of interview
Here is there and last part of my interview with him:
Anmol Rajpurohit: Q8. What resources would you recommend to learn and practice TDA? Is Topology adequately covered in Data Science related educational curriculum across universities?
Anthony Bak: TDA is not currently part of any standard data science curricula, but there are a growing number of schools that offer graduate level courses in TDA. Penn, Duke, Stanford, University of Chicago, University of Ohio all come to mind, but there are doubtless others. The perspectives and techniques are not exactly what we do at Ayasdi, but there are many shared ideas. A nice example syllabus with resources: http://web.cse.ohiostate.edu/~tamaldey/course/CTDA/CTDA.html
For self study I also recommend starting with "Topology and Data" by Gunnar Carlsson (http://www.ams.org/journals/bull/
20094602/S027309790901249X/S027309790901249X.pdf). If you're interested in applications to Machine Learning or particular industrial problems, I recommend contacting Ayasdi and asking for the relevant white papers as there is relatively little published in the academic literature. For a more gentle introduction for a working data scientist I have a number of video lectures up on YouTube that I've given at various meetups, academic seminars and industry conferences (sorry for the self promotion).
AR: Q9. What motivated you to work in Data Science?
AB: As a researcher I worked on connections between geometry and string theory, and while I found string theory satisfying in an abstract sense I wanted to work on projects that had more immediate real world impact. I started looking outside of academia, and data science was a draw for a variety of reasons.
I was particularly drawn to the vibrant culture of creativity in data science; it pulls from whatever discipline it needs to without dogma. Data scientists themselves are constantly seeking novel ways to quantify and explain the world around them, and while some people argue about the term "Data Scientist" and say they are really statisticians, analysts, or machine learners, I think that culturally we are distinct and embody the scientific approach to the world in ways that others do not.
AR: Q10. What is the best advice you have got in your career?
AB: "Work it out for yourself. Start with a simple example."
AR: Q11. What key qualities do you look for when interviewing for Data Science related positions on your team?
AB: We look for people with some quantitative background in mathematics or the natural sciences. At Ayasdi, Data Scientists work with customers to solve their problems, do product development, and more fundamental R&D. This is an extremely broad set of responsibilities and we generally expect people to be solid contributors in two of the three areas.
R&D is more important for us than for many data scientists embedded in industry. We work at the cutting edge of TDA and machine learning and can't necessarily rely on academic literature to find solutions for our problems. Some of what we do at Ayasdi challenges machine learning and statistical orthodoxy, and we need people who can independently and creatively think through our methodology and distinguish a heuristic from a theorem.
One thing that we require is that all Data Scientists, even those with a more research oriented bent, work with customers. This keeps all of our efforts focused on solving problems as they appear in the real world and not in idealized settings where something can be proved. As a result, communication with both technical and nontechnical counterparts at other companies is a requirement for a Data Science position at Ayasdi. We ask all interviewees to give a seminar type presentation on a data problem that they've work on. The presentation is evaluated for both content and clarity.
Much harder to define, but something that we look for, is an "intuition for data." This is a combination of quantitative skills, broad curiosity across a variety of disciplines, creative transfer between disciplines, a nose on where to go look for an answer, and the sense when you've found something interesting.
PM me if you have a good way to measure this in an interview.
AR: Q12. On a personal note, we are curious to know what keeps you busy when you are away from work?
AB: I like to explore the Bay Area with my family. While job opportunities in Tech is what brought me here, the cultural and culinary diversity set in stunning physical beauty is what keeps me here.
Related:
Top Stories Past 30 Days

