How to prepare for a career in Data Science

Slides from my Learning Workshop session at Udacity Intersect Conference 2018.

It takes more than a resume and an interview to build a successful data science career. You need to have a solid understanding of your goals—both short and long-term—and a clear vision for the path you want to take. In this session, you'll learn how to find the role that’s right for you. We’ll discuss best practices for your resume, how to prepare for a successful interview, and the importance of building your portfolio. We’ll even look at accepting the right offer, and what to expect at your first job.

Transcript

1. How to prepare for a job in Data Science Praveen Gollakota @pgollakota Director of Data, Udacity 2018-03-27
2. Who is this for? • Anyone looking to break into the data science field. • More catered towards someone looking for their first job 2
3. The agenda • Understand the industry • Understand yourself • Understand the job • The interview and the offer • What to expect on your first job • Planning your career • Odds and ends • Fin 3
4. The Job Market A quick survey of 100+ data science jobs
5. What employers are looking for 5
6. What employers are looking for 6
7. What employers are looking for 7
8. Data Scientist is the sexiest job of the 21st Century - Someone who really wants someone else to deal with the messy data
9. Job Realities 9 Most of the jobs are heavy on ETL, reporting and visualization. Know your SQL
10. Skills 10 ETL - Extract Transform Load SQL - Structured Query Language Visualization Modeling
11. Type A: Analysis Making sense of data or working with it in a fairly static way. Statistics, Data cleaning, methods for dealing with very large data sets, visualization, deep knowledge of a particular domain, writing well about data, and so on What is your type? Type B: Builder Interested in using data "in production." Strong coders and may be trained software engineers. Build models which interact with users in production. 11 This classification of data scientist roles comes from Michael Hochster’s Quora answer for “What is Data Science”
12. Applying for a job
13. The right company Small ETL & Reporting Metrics No process Self-directed Medium A/B testing Reporting Analysis Product iteration Large Data Products Recommendations Scale 13
14. Resume 14 • Customize for size of company • Customize for Type A vs Type B • Customize for your previous experience • Keep to one page • Emphasize data prep, SQL, etc.
15. Interviewing Company, Resume, Interview Prep
16. What to learn 16 • Data munging, SQL, Pandas/R. • Visualization - ggplot/pandas/plotly/etc. • Programming - loading files, reading from APIs, etc. • Linear methods, Random Forests • Preprocessing & model selection
17. Know the company 17 • Do your research • Research the size, projects, tech stack and more • Think about the problems the company may be facing • Have an opinion. Will data science help at this stage? • Does a recommendation algorithm help their bottomline? • Know why you’re here
18. What to ask • What is the most recent project? How long did it take? • What was the main challenge? ETL? Framing the problem? Productionizing? Selling internally? Justifying the cost? • What are the current team priorities? How often do they change? • How do they prepare the data? • Who is the go to expert for various topics on the team? What about engineering? • What did they learn in the job? • What was their impact? 18
19. Interviewing Realities 19 • Interview process is imperfect • Data scientist interviews vary a lot • Don’t be discouraged
20. How to choose Which offer is the right one?
21. Choosing the one 21 • People, people, people • Diversity matters (of thought, of experience, of life) • Do you want to hone your craft? • Do you want to make an impact? • Does it help with your long term career goal?
22. Your First Job
23. What to expect 23 • Data won’t be available • Documentation is going to be sparse • Job role won’t be well defined • Communication is key • Know what you want • Jump in and contribute
24. Being successful 24 • Don’t lose focus of your long term goals • Everybody has an unconscious bias. • Be diplomatic. Pick your battles. • Learn everyday • Have realistic expectations • Do things that matter
25. On Impact
26. On Impact 26 • Optimize for long term impact • Know what you can impact • Sometimes data science is not the right tool for maximizing impact. Do a back of the envelope calculation • Be honest with yourself about it.
27. On Data Science
28. Cargo Cult Data Science 28 • Data Science is a hodgepodge of various disciplines. There are a lot of practices and incantations from one discipline that don’t apply to another • Test everything empirically. • Learn how to simulate anything quickly. • Don’t let jargon intimidate you.
29. Endnotes
30. Your Goal Find the job that’s best for you 30
31. “ Best Job = max f ( Who you are, Who they are, Where you are, Where you want to go, What you want, What they want ) 31
32. The variables Who you are Education Experience Values Beliefs What you want Experience/Knowledge Exposure Team Brand/Money Where you are New job Career change Industry change 32 Who they are Size of the company Industry Team Where you want to go Start a company Product Management Management Researcher/Technical Expert Software Engineer What they want Experience level Team fit Skills
33. The best job for your friend is not the best job for you 33
34. Know what you want 34
35. 35 Thanks! Best of luck! Any questions? @pgollakota