Kaggle Career Con 2018: Day 1 Recap

Kaggle Career Con 2018 launched today at noon PT, commencing a 3 day event (3/20/2018 - 3/22/2018) aimed at helping attendees land a first job in data science or transition into data science from another field. Day 1 consisted of 3 live streams which are available on Kaggle's YouTube Channel. This post aims to give a brief review and recap of the key takeaways of each session.


Session #1 Recap: How to Build a Compelling Data Science Portfolio and Resume

The first talk given by William Chen, a Data Science Manager at Quora, covered dos and don'ts of a data science resume. Chen broke his talk up into a list of 9 key points that he recommends for constructing an effective data science resume:

1. Keep your resume to 1 page - Hiring managers don't have time to look at more. Cover letters are optional and only really useful if you can tailor it to the company.
2. Don't include an objective section - They are generic and waste space.
3. List relevant coursework - Any experience related to the position should be highlighted.
4. Don't give numeric skill ratings - They have no standardization and are essentially meaningless.
5. Do list technical skills - Particularly those mentioned by or directly relevant to the position.
6. Don't list common projects or homework - For example, avoid listing learning competitions on Kaggle like working with the Titanic dataset or MNIST as half of the data science community has worked on them.
7. Do show results of more involved projects and include links - If you scored highly on a featured Kaggle competition (top ~10%) that could be worth mentioning.
8. Do fill out your online presence -  Profiles on sites like LinkedIn and Kaggle as well as other online presence like a website or blog can help managers find out more about you.
9. Do tailor your listed experience for the job.

The key takeaway of the talk was that you should keep your resume clean and concise and tweak it to be as applicable as possible for the actual job you are applying to. Even something as simple as changing the order in which you list your skills so that skills mentioned in the job description come first in your lists can make a difference. Overall it was a very good talk with a lot of actionable advice, although it did not really spend much time discussing portfolios and a lot of the advice assumes that you are applying to a specific position rather than submitting a resume to a general pool or job board.


Session #2 Recap: Am I a Good Fit? Identifying Your Best Data Science Job Opportunities

The second presentation by Jessica Kirkpatrick from Slack focused on differentiating between different types of data science roles and assessing which ones fit your skillset. The presentation focused on the types of skills, experience and responsibilities for 5 different data science roles:

1. Data Analysts - Create insight from data. Requires a mixture of data skills and soft skills.
2. Data Engineers - Build tools for analyzing, visualizing and sharing data within the organization. Requires strong programming skills.
3. Data Architects - Design systems to get and store raw data and make it accessible. Requires strong computer science, programming and database skills.
4. Domain Experts - Data scientists with experience in specific domains. Typically requires a PHD or extensive industry experience.
5. Managers - Manage data science teams. Ideally they have worked in other data science roles preciously so that they understand and can assist with the technical side of projects.

The second part of the talk discussed various data science position seniority levels. In order of ascending seniority and knowledge:

1. Data/Business Analysts - Junior level data scientists; do not typically require a lot of advanced technical knowledge or an advanced degree.
2. Data Scientist/Senior Analyst - Usually requires and advanced degree or significant work experience and knowledge of programming and statistics.
3. Senior Data Scientist: 3+ years of experience in the same industry.
4. Staff Scientist: The most senior contributor on a team.
5. Lead Data Scientist - Management role that requires both technical knowledge of a staff scientist and management skills.
6. Manager/Director/VP/Head of Data Science - Senior level management role.

The talk did a fine job laying out the types of data science job descriptions you'll see and the sorts of skills that might be required for them.  Since the convention is for people looking for their first data science job, I'm not sure how relevant high seniority positions are to attendees, but it does give an idea of the sorts of positions you could advance to after gaining experience. The session didn't contain a lot of directly actionable advice, but there was one key takeaway prompted by a user question that is worth repeating: it is not your responsibility to determine whether you are qualified for a position, so don't be afraid to apply to positions for which you do not meet all of the stated requirements. Sometimes requirements are flexible.


Session #3 Fireside Chat with Dr. Fei-Fei Li

The final session of the day was an interview-style live stream between Kaggle CEO Anthony Goldbloom and Fei-Fei Li, Director of Stanford University’s AI Lab and Chief Scientist at Google Cloud AI/ML of ImageNet fame. The session highlighted on Dr. Li's personal story as a 16 year old Chinese immigrant who fell in love with physics and ultimately neuroscience and AI. As a story-driven talk, there wasn't a lot of actionable advice for job seekers in the session: it was more about inspiration and the big picture of AI. The key takeaway was that you should give yourself the freedom to explore and follow your interests: there is not one "right path" to get into data science or secure a job at an elite company although a PHD is probably the best way to get involved in AI RND.