An Inside Look into Mentorship at Thinkful

Last week, a new data science mentor asked her peers for some advice on being an effective mentor. Rohan Kekatpure, a Data Scientist at Intuit and Stanford PhD, responded brilliantly with a few emerging themes from his own sessions. This thread is a preview of the insightful converations that take place within our education and mentor teams. It’s time for us to make them public.

SUBJECT: Tips for a new mentor

I am starting my first student sessions tomorrow. Given that the course has been going on for more than a month now, I was hoping that some of you would have some tips for me from your experience.

I will greatly appreciate feedback on what topics students have been having issues with and any other tips in general regarding the mentoring sessions.

RESPONSE from Rohan

Welcome to the team! The January batch is my first remote mentoring experience as well. So the following will change as I progress further along in my journey. Regardless, some themes have indeed emerged from my first couple of sessions and I’d like to share them:

1. Why mentors?

Thinkful course materials are well organized to answer the how part of doing data science (how to connect to a database, how to instantiate a Pandas data frame, how to train a model). Mentoring aspect, in my mind, should fill in all other parts (why do we need a database in the first place, what role does a database perform in enterprise software? why do we need Pandas if we already have databases? when should you use logistic regression vs SVMs). No structured course, Thinkful included, can reasonably be expected to fill this gap using static course materials alone. This is where we mentors come in.

2. Start each session with a high-level overview

I usually start off my weekly sessions with a quick 5-10 min high-level overview of the topic, my experience with it and where it fits within the framework of professional data science. I started it as an experiment and have observed that it serves two aspects: (1) Students immediately see how their weeks’ learnings are relevant professionally and (2) this perspective inspires them to understand things at a more fundamental level (e.g., why does Python do it this way? whats the underlying pattern in this workflow?) This is a part of my ongoing attempt to provide purpose for the week’s course materials. 

3. Be the student

Before my weekly session, I code-complete all things that I anticipate we’ll cover in our weekly session. I go through the lesson pretending to be the student. I stress code-completion vs reading for two reasons: (1) My conversations with my students occur at a deeper level if I’ve pre-experienced their journey through the lesson and (2) I’m able to better address Point # 1 above by doing pre-research (Wikipedia, Stackoverflow, textbooks etc) on our week’s topics. After the student has had a chance to work on the week’s exercises, share your code with the student. Reading good code is a great accelerant for students’ learning process. 

4. Plan ahead

Reach a mutual agreement with the student regarding what you’re going to discuss next week. This is suggested in the mentor field guide, but I discovered its effectiveness only after I began practicing it. 

5. Get feedback and realign

Every two weeks, I ask my students if they find our sessions satisfying. This gives students a chance to provide a feedback on YOU and Thinkful. I (plan to) do this often and not as a one-off thing at the course-end. This will, I believe, provide me a chance to re-align the course of the ship.

6. Will you hire them?

Most students ultimately are seeking career change to data science. So I pretend that they’re an intern where I currently work and ask myself, “Will I hire this person as a data scientist in my team?” If the answer is negative, I ask “Ok, what skills do they need to be a strong candidate?” I then think about how to fill their existing skill gaps using a combination of Thinkful materials, internet, my experience and my understanding of the big data landscape. 

Please feel free to discard or dispute any of the above points. Ultimately, this is a personal journey for everyone. However, I do welcome any viewpoints that will allow me to be more effective with my students. 

To learn more about our Data Science course, visit the course page.

If you’re interested in joining an incredible team of engineers, designers, and data scientists, you can apply here