Skip to main content Skip to secondary navigation

Getting a job in DS/ML and the things that matter the most once you’re in that no one tells you about.

Main content start

Getting in, and getting ahead. What are the best ways to start your career in data science and/or machine learning? And how do you stand out and get promoted once you’re in? Paul Magon de la Villehuchet, MS ’19, hosted a session full of practical advice about tailoring your resume, what to expect in interviews, and the skills that make the difference to be impactful.

About Paul:

  • Originally from France, undergraduate degree is math and physics.
  • MS, MS&E 2017 – 19, took a range of classes from entrepreneurship, hardcore math, AI/ML in CS
  • Interned at Uber, twice, in data science
  • Work experience 
    • QuantCo for a year, during Covid. 
    • Moved to Upstart in ML verifications team: connecting financial market and banks to people who need personal loans (and soon, mortgages). 
    • Started there as research scientist, now manager.

 Why Paul enjoys working work in ML:

  • Holy trinity of work – interesting, impactful, well paid. 
  • ML creates a lot of value at companies, so important decisions are being made in this area. 
  • It is a field that is evolving quickly, so on the cutting edge of things. 

What are ML companies looking for?

  • Venn diagram – product, engineering, stats.
  • ML is the intersection of the product, engineering and stats. Coding is important.
  • Pay attention to job titles and keep an open mind: different companies have very different titles that map to the same thing, for example: data scientist/data analyst/ML engineer: all the same titles though!
    • So don’t pay too much attention to the the job title, look at the job descriptions.
    • Highlight your skills/duties/responsibilities on your resume so the recruiters and hiring managers can map what you have to their job titles and responsibilities.
    • Since so many titles can mean the same thing, and might be different to what they call the position at their company.

How to stand out in an interview:

  • ML interview – there are 3 things that will be assessed, and the weighting will depend on the role and company, so these are not listed in any particular order.
    • Engineering – practice leetcodes and ML case studies. 
      • Eg, taking a dataset and building a model.
    • Stats and quant skills – brainteasers, regression problems, ML case study (also covers product). 
      • Eg, classification problems.
      • Pay attention to where you’re fast, and where you can improve. It’s better to do something simple and well, in a short amount of time, than something overly complicated for the sake of impressing them.
    • Culture – questions around whether you fit or not.
      • Surprising number of people don’t pass this part of the interview. 
      • Each company has their own culture, eg, collaboration, work-life balance. Be honest about who you are.

General career advice:

  • Be the master of your own destiny!
    • Optimize for the long term. If you’re in the wrong place, then leave, but ideally you can take the time to build trust within your company. This trust gives you more freedom and autonomy to explore areas and experiment.
    • If you’re wanting to develop and grow your career, you need to focus on self improvement.
    • Be as close to the value creation and impactful projects within your company as possible. 
    • Do what you’re supposed to do, and do it extremely well. Too many people are overly ambitious and focus on the job they want to do, rather than what they’re getting paid to do. Once you’ve established a good reputation and shown your reliability, that's when to start expanding your scope beyond what you’re supposed to do. Successful people start by doing something well, have a little luck, and take on more responsibility. This is what will make you stand out.
  • Good companies should notice talent and reward it, but you still want to make sure you’re on track. Ask for feedback!
    • Create a culture for yourself of asking for feedback. Paul loves it when people give him feedback because then he can improve.
    • When you get feedback from others, be graceful (it can be as hard to give it as it is to hear it!), and then act on it.
  • Think about what everyone cares about. What the company cares about, what your team cares about, what you care about rarely all align. So this is where things can get hard! But if you pay attention, you’re more likely to make more people happy!
  • Figure out how to make an impact and create long term value for the company. This can be hard to measure, but great if you can demonstrate.
  • Focus on your skills:
    • Your tech skills, eg coding, stats, are objective and therefore easy to measure, but they’re not the most important! (That’d be the next things listed).
    • Your technical judgement/ability to make good decisions. In reality, you’re winging it a lot of the time, so your intuition is important. This takes time to learn, because it develops with experience. And it takes time for others to learn to trust you. As you take on more responsibility, this becomes more and more important.
    • Ownership – your ability to reliably and proactively take on responsibility.
    • Communication and organization – your ability to work well in a team. Communications with non-technical audiences are important here too.
  • Be strategic: 
    • Optimize for simplicity. Do the simple things well, before moving onto the more advanced things. Identify the most valuable thing, do it, move on to next.
    • Optimize for iteration time – ML is annoying sometimes in how long it can take! A model can take a day to train, and bugs will slow you down. So, start simple, de-bug, and then move on to the next model. Test smaller data sets, de-bug, then run the model again with the full data set.
  • Be curious. Read/learn about the other teams and wider topics. Get to know people on other teams. This improves your ability to generalize and your overall knowledge of your company and the industry.
  • Be autonomous - eventually. Early in your career, ask questions and get help when needed. After a while, you shouldn’t be asking basic questions any more.

 Convinced this is the career path for you and wanting to get started?

  • Upstart is hiring! They’re looking for new and recent graduates for full time roles, and also have internship opportunities. Contact Lindsey Akin for an email of introduction.