As I write this it is Christmas Day, 2015.   I am working as a Data Scientist at Payoff in Irvine, CA, but last year I was a physics teacher.  It took a lot to make the change.  I wanted to take the time to reflect and share that transition that started in January of 2014 when I started taking my first course in the Online Masters of Computer Science (OMSCS) with the goal of leaving teaching and becoming a data scientist.  

How I started and stopped Teaching

I started teaching at The Buckley School (private school in Sherman Oaks) in Aug of 2007. I accepted the job impulsively right after finishing up my Ph.D in Theoretical Particle Physics and Cosmology.   The idea of job insecurity as a postdoc for 3 to 9 years, working on topics that no one cared about (or understood), and having to leave my then girlfriend/now wife was too much. I received an email about an open teaching position that offered me a chance of something new and to stay with my now wife. So I went in for an interview, and they offered me the job before I left the campus.

The one thing that surprised me about teaching was how difficult and exhausting it was.  I was also surprised with how rewarding it was.  The problem was that the exhaustion was winning out in the long run.   By 2012, the school’s politics taxed me.   It was a high-end school where questions like the following were posed to me:

How can I ask Mrs. XXX to donate money for our new building phase while her daughter has a B- in physics?

It seems like a bad line from a movie, but it stems from real concerns in institutions that rely heavily on donations to execute their mission.  It is leadership’s role to strike the balance in the relationship between giver and receiver.  Every leadership team will take a difference stance.

This event motivated to seek to renew joy by switching to a school that was more aligned with a social mission and making a difference in a larger community.  This is when I transitioned to New Roads.  This place was more wonderful, more meaningful, but more exhausting than Buckley.  It was here that I finally burned out and needed a different career.   

My PhD in Physics was under-utilized in teaching, and I wanted to take advantage of these skills.  I enjoyed executing on these skills.  I also wanted to work on developing insights (like in graduate school) that lead to actions (unlike in graduate school).  Somewhere online I found this mythical job called “Data Scientist”, and I got a tingling all over my body:  This is it!

Choosing to Become a Data Scientist

I found my calling, but in my mind it was not as easy as applying for a data science job.  The reason is that I categorized the skills/requirements for the positions I found online into a couple of buckets:

  1. What is that?
  2. I’ve read about that.
  3. I’ve played with that.
  4. I’ve done a project involving that.
  5. I’ve done multiple projects involving that.
  6. I’ve lived and breathed that.

Most of the buckets were 1, 2, & 3.   I know other PhDs who would have made the argument that they could learn it–they have a PhD after all.  The issue to me is not the truth of this statement, but the relevancy of it. Academia hires you for what you can learn, industry hires you for what you can do.   I did not want to find myself in a position where I can not deliver on what was needed, or only get hired by people who can not adequately decide that I am not the right person for the job.   Fit was and continues to be one of my major concerns for employment.  I want my skills and abilities to fit with the need of the company that hires me.

So in January of 2014 I took the following actions to facilitate my transition to becoming a data scientist:

  1. Enrolled in OMSCS
    1. Machine Learning
    2. Artificial Intelligence for Robotics.  
  2. Put our house up for sale
  3. Applying for teaching jobs near my in-laws
  4. Plan to live with my in-laws for at least 1 year
  5. Commit to being unemployed starting the Summer of 2015.

The Long Slog

That this point the story becomes uninteresting because it is me just putting my nose down for months on end, working on anything that gets me to produce deliverables.   I found for myself that I learn best when I have to give something to someone.  That’s actually the reason for this blog.  

What I did was integrate this work in my life.  I committed every weekday morning from 4:30am to 6:30am working on data science or masters in computer science related work.   I also committed most Saturday morning and all day Sundays to this work as well. It is not hyperbole to say I spent more than 20 hours/week working towards this transition.  I also reduced my social life quite a bit during this time.  That probably goes without saying.

Job Applications

I started applying for data science jobs in January of 2015.  At first I received zero responses, so I hired a freelancer that worked for as a HR person at a tech company to re-write my resume.  She did an excellent job.   I immediately started getting responses after using her version.  Recruiters (who now I will not work with) started reaching out to me.   It turns out there is a keyword and format game that must be used to get results.  It also turns out that you do not want to say you were a  teacher.  My teacher-less resumes received more traction.  All in all I applied to well over 100 jobs, did dozens of phone interviews, and had 10’s of on-site interviews.  

I learned that start-up companies are the quickest to respond and the quickest to dismiss.   I learned the established companies can take up to 3 months to get back to you, and sometimes that has nothing to do with you or their interest in you.  This does not seem to be true in San Francisco, however.   People are very quick to respond if interested.   


In the winter of 2014 Udacity started their nano-degrees.  There was one for Data Science (which was later changed to Data Analyst).  It was perfectly aligned with my goals, project based learning focused on Data Science.  It was affordable, online, using the same platform as OMSCS.  For me it was a no-brainer.

The value-added for me were the projects and feedback.   They were opened ended, so you had the option to do as much or as little as you wanted.  There was also very good feedback associated with each submission.  My one recommendation is that if you use Udacity, invest time into summarizing them into a pitch deck that can be talked about in less than 5 minutes.  This can be useful to show off to potential employers, and much better than showing them a 20+page report.  Plus it has the added benefit of working on your communication skills.

This also lead to the best decision I made involving my career transition.  The 5th part of the nanodegree is Data Visualization with D3.js.  If you click on the link you will see that it was built with Zipfian Academy.  I was curious about them, who they are, what they did.   It turned out they were a data science boot-camp in San Francisco that was acquired by Galvanize.  After reading about them on Quora, I decided to apply.  My job search was not going well at this point.

Galvanize Data Science Immersive

The application process for Galvanize was  a series of programming and statistics related questions and interview.   The boot-camp is immersive, and the more prepared you are the more you will get out of it.  It is also a financial commitment.  The tuition was 16k when I went, and after living in San Francisco for 3 months I was out about 24k.

After I was accepted into the program, but before I committed, I was offered a Data Science job.  This was a pivotal moment for me because I was being offered my goal/dream.   I could accept the job, not spend the money at Galvanize, would have to move, but work as a full time Data Scientist.  The problem for me was that it didn’t feel right.  It was not a ‘hell yes!’ decision.  I also did not feel like I would be accepting the position from a place to strength or with a full knowledge that fit into the position.  I also did not align with the mission of the company.   I thankfully opted for Galvanize.

The Galvanize Data Science Immersive was the hardest fun I ever had.  More importantly it put almost all the data science skills/abilities I started to list in 2014 into the 5 & 6 categories from earlier.  I was also, in Seth Godin terminology, shown my tribe.  It allowed me to build a network that will continue to help me professionally, gave me exposures to a wide rage of companies and interviews, and allowed me to know for certain if I and a given position was a correct fit.  I can not give this program enough credit transforming me into a high quality Data Scientist.


I accepted a job at Payoff in August of this year, a full 20 months after I decided to become a data scientist.   There were other options and different paths I could have taken, but I think the path I took was near optimal in terms of long term rewards.  I get to work at a company whose mission is being a financial institution that helps people transition from being borrowers to being investors.  Science and Compassion are integrated into Payoff, so we get a treasure trove of data with a mission to find actionable insights that lead to making people’s (financial) lives better.   And because of the choices I made I have learned and mastered the data science skills necessary to contribute on a day to day basis.


It was a difficult two years, with a number of big and stressful decisions.   All in all I spent 31k transitioning from teaching to becoming a data scientists.  I am happy to say that this will be recouped in less than two years (even after taxes) in the difference between what I was making and am currently making.  Also, I go to work everyday looking forward to what I am doing.   I do not think you can put a price on that.

There were dozens of times I wondered if it was worth it, or thought about just teaching.  It was a certain, stable, and noble profession.  I am glad I stuck with it because I know too many teachers who continued teaching after they lost the passion.   It’s not good for anyone.

It took about 2 years, but both my wife and I are happier than we ever have been.  We are glad that I decided to take the long hard road.

Thank You

I appreciate you taking the time to read this post.  I hope you gained something from it.  Feel free to reach out with any questions in the comments or directly through the email link in the menu.  


Join the Conversation


  1. Hi Bryan, thank you for the post. I am in a similar position, transitioning from academia to data science. I’ve been referencing your “Resources” page and find it very helpful. I was wondering if you could also list the books that used in your journey?


Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: