How Data Science Can Nearly Double Your Income, with Zachary Washam

In data science, knowledge is not power… applied knowledge is power.

Businesses want to know if you can deliver results that impact the bottom line. But how exactly can you do so? Are hard skills enough? What practical steps can you take to help your organization thrive?

We explore these questions and much more in this interview with Zachary Washam, an investment banker leading a data science team at Wells Fargo Securities.

Zach was a student in one of the first cohorts of our Machine Learning Masterclass, and he recently reached out to share what he’s been up to since then… His story was too inspiring not to pass along!

Over the last year, Zach was able to:

  • Learn data science while working a full-time job.
  • Successfully pitch a ML project to the head of his bank.
  • Be appointed the project manager.
  • Recruit an international team of statisticians and data scientists.
  • Land his name on a patent as an inventor.
  • Become nominated for a promotion that would nearly double his income.

We break down exactly how he did it, share actionable tips and takeaways for others, and discuss the future of machine learning in financial services.

So without further ado… let’s get to the interview!

How Data Science Can Double Your Income with Zachary Washam

Hi Zach, thanks for being here! Could you please introduce yourself and describe what you do?

I am an investment banking analyst at Wells Fargo Securities specializing in debt capital markets. I work with large, investment grade clients (Wal-Mart, Verizon, Delta Air Lines, etc.) to raise money for acquisitions and operating liquidity.

It's a highly analytical role involving large amounts of market data, and there is huge untapped potential to leverage this data for useful business insights.

How do you currently apply machine learning in your work?

Using what I learned from EDS, I pitched the head of the investment bank for sign-off to build an analytical tool that predicts investor behavior. With his approval, I was appointed project manager and subsequently recruited a small international team of data scientists and statisticians.

Over 12 months, we successfully developed Wells Fargo Securities' first-ever debt capital markets machine learning algorithm. We have engaged legal counsel and are patenting the algorithm, and I will be listed with two members of my team as one of the inventors on the patent.

That’s incredible to hear! Could you tell us more about the debt capital markets machine learning algorithm?

The algorithm leverages proprietary transaction data - pricing, structure, company financial data, etc. - and identifies relationships between these input features and historical investor behavior.

Once the algorithm is trained, the user can input hypothetical transaction details and the algorithm predicts investor interest in terms of how much capital they are likely to contribute.

Since Wells Fargo Securities is one of the top three banks in this area, we have more proprietary data than almost all of our competitors. My goal is for this algorithm to help us deliver superior insights to clients and retain a leadership position in our market.

Do you plan to tackle other challenges with this team of data scientists and statisticians that you’ve assembled?

Absolutely - the results of our first algorithm have inspired the whole team and generated significant buzz within the company. We have access to a massive amount of useful data in our business, and there is no shortage of applications for machine learning.

We have already started on a second algorithm and hope to finish in the first or second quarter of 2018, and I doubt we will stop there.

This project sounds very impactful for your bank. Has it also influenced your personal career?

Yes, the success of this project has super-charged my career trajectory. After less than two years in my position, I was nominated for a promotion historically restricted to those with graduate degrees (which I do not have) that would nearly double my income.

I now have regular dialogue with senior leaders across the investment bank, including the head of the investment bank, and I have been pegged as a high performer with a bright future.

Many of our readers are in a similar situation as you were in 12 months ago - smart, motivated, and looking to take the next step - Could you break down how you proposed this project and share any tips for getting these types of projects approved?

About a year ago, the head of the investment bank spoke at a company event I attended. During a Q&A session after his speech, I asked him about our machine learning efforts, and he appeared to be interested in the topic. I approached him afterward and asked for 30 minutes to pitch him my ideas, and luckily, he decided to squeeze me into his schedule a few weeks later.

I think the presentation resulted in project approval because I was able to clearly quantify the business value of my proposal, and with low risk and significant potential to boost revenue, it was kind of a no-brainer.

Besides expertise in debt capital markets and hard skills in machine learning, what’s the biggest factor that has helped open these opportunities for you?

You have to be able to sell your vision, which requires speaking confidently and concisely. If you get the chance to speak with a senior leader like I did, you probably only have about 30 seconds to make your point. You want to communicate that you are capable, understand their priorities, and can deliver a solution with real business value.

If you're recruiting a team - which was key for me, because I couldn't have done this on my own - the same rules apply. If a potential partner is super talented, they will have other opportunities to take on projects. For them to buy-in, they need to believe that you are going to succeed and that the project will have a real impact.

It seems that many major financial institutions are investing in machine learning. How do you see ML’s role in your field evolving?

Machine learning is transformational for investment banking. Right now, most ML applications in banking, like trading algorithms, are trained with public data, because there is a lot of it and it is high quality. The problem with using public data is that it does not give you a sustainable advantage. Since the data is public, your competitors will eventually replicate any valuable analysis you produce.

Over time, I expect that the development of machine learning capabilities will drive banks to focus on generating a greater quantity of proprietary data to feed their algorithms.

For example, U.S. banks that specialize in public domestic transactions may become more interested in playing in the European market, which is much more opaque and provides greater proprietary data for market leaders.

When did you first become interested in machine learning? Was there any particular challenge or idea that prompted you to start studying this field?

A big complaint from bankers in debt capital markets is that clients view our product as a commodity, as if there is no significant performance difference between the top handful of banks.

As I was thinking about ways to differentiate my company from our peers, I noticed that we had a large amount of proprietary data due to our historical market leadership. Searching for methods to leverage this proprietary data led me to machine learning.

At the beginning, what did you find most confusing about machine learning?

Coding - I had never written a line of code in my life before taking on this project. Luckily, Python is relatively straightforward compared to other languages, and I had some super intelligent (and patient) partners to help me out.

Were there any “aha” moments during your studies that helped things “click?” What would you recommend beginners to do in order to get caught up to speed ASAP?

Honestly, it was more of a slow and steady process for me, and consistent hard work was critical. Even when I was in the office past midnight, I tried to study or work on the project at least 15 minutes when I got home every weekday.

On weekends, I'd grab my laptop and hunker down in a coffee shop for hours. Sometimes I'd be there so long, they'd stop giving me free refills!

So far, what has been the most useful or important ML concept for your work?

Good data beats fancy algorithms.

Zach, thanks so much for sharing your story with us. It’s inspiring for newcomers and veteran data scientists alike. Final question! If you could give just ONE piece of advice to others who wish to accelerate their careers with data science and ML skills, what would it be?

Target goals that are out of your comfort zone, and trust yourself to figure things out. To paraphrase Steve Jobs, everything around you was made by people no smarter than you, and if you believe that you can make things better, you probably can.

Zachary Washam is an investment banker leading a team of international data scientists at Wells Fargo Securities.

Final thoughts from EDS: 

Hearing Zach’s story reminded us of a quote from Richard Branson: “Chance favors the prepared mind. The more you practice, the luckier you get.”

In Zach’s case, he was already well prepared before the chance arose to pitch an idea to the head of his bank:

  1. He had been diligently studying coding and Python at nights and weekends.
  2. He had learned enough about machine learning to understand its capabilities.
  3. He had already thoughtfully considered how his company could differentiate itself with its proprietary data.

We are only seeing the beginning of the opportunities to leverage machine learning to make an impact (and accelerate your career). Will you be prepared for them when they come?

Comments are closed.