5 Steps to Launch a Successful Data Science Summer Intern Program

Starting a data science summer internship program from scratch can seem daunting. How do you ensure you hire the right people, enable them to succeed, empower employees to develop professional skills and have fun? When is the right time to start an intern program and how can you measure its effectiveness? In this blog, we’ll walk through the 5 steps ShopRunner developed for building and executing a successful data science summer internship program.

Before we begin, let’s first define the purpose for our program:

• Extend talent acquisition pipelines for future data science hires
• Make an offer if we found a great intern
• Increase ShopRunner brand awareness
• Develop management and professional experience for current employees
• Increase team diversity
• Have a ton of fun!

Next, why did we decide to start an internship program now? ShopRunner, though still a small startup, is an innovative and rapidly growing company. The data science team, in particular, had demonstrated value and was looking to expand by hiring junior folks.

We will now focus on the 5 steps we developed for launching our program. They include:

• Hiring
• Planning
• Managing
• Fun
• Retrospective

1.0 Hiring

How many interns should we hire? We based this decision on two factors:

1. How many teammates wanted to mentor an intern
2. How many interns our company could financially support.

Though all five data scientists on our team wanted to participate in the summer internship program, we were able to allocate funding for two interns. We designed our recruitment process to hire this count and began by identifying a roadmap and timeline for the process. Our target start date for the program was Mid May, 2019.

Timeline

• December, 2018 (5.5 months out): Ideation and purpose definition for the internship program
• Mid December, 2018 (5 months out): Internship approval by Engineering, Finance, HR
• Mid January, 2019 (4 months out): Internship job req posted and marketed
• February, 2019 (3.5 months out): Resume and phone screens
• Mid February, 2019 (3 months out): Final round interviews
• March, 2019 (2.5 months out): Offers out
• Mid March, 2019 (2 months out): Offers accepted
• Mid May, 2019: Program Start

Funnel

To hire two interns, we worked backwards with our HR team to estimate the number of interviews, phone screens and resumes we’d need to sort through. We designed a funnel which looked like:

• 150 resumes screened
• 20 phone screens
• 4 final rounds
• 2 offers, 1 backup
• 2 hires

During resume screening, we focused on candidates who were currently studying or researching in a relevant field, had python experience, and were hirable within a year of the summer internship. Phone screens helped identify a candidate’s interests, clarify their background, and understand their communication skills. They were critical in screening out candidates not yet ready for final rounds.

Prior to the final round interviews, we asked candidates to either a) send us a code sample or b) complete a short take-home exam. Interview #1 of the final round reviewed a candidate’s project code and assessed if the he/she had the technical chops to be successful. Interview #2 focused on a real business problem ShopRunner faced and assessed if the candidate could logically reason about the use of machine learning to help solve the problem. Finally, interview #3 assessed the personality, background, and fit of each candidate to ensure he/she was aligned with our company’s values.

We found four final round interviews was enough to extend two good offers. After our offers went out, both interns accepted! It was motivating, exciting, and invigorating for our team to have come together and bring in two new members for the summer!

2.0 Planning

Planning for a new program is critical, but challenging. We began reading online content to learn from others’ successes and pitfalls. Some of the more helpful posts are below:

• So You’ve Hired Interns, Now What? How to Thoughtfully Plan a Summer Intern Program
• How to Build a Successful Internship Program
• Airtame’s 4 steps to a successful internship program
• Creating A Successful Internship Program In 6 Easy Steps
• How to design an effective startup internship program

From these articles, we identified the following points as being absolutely critical for our program:

• Interns would help select their summer project from predefined list
• Onboarding document to provide guidance during the first week
• Interns assigned a running buddy, a friendly peer to help them ramp up
• Weekly feedback and a 360 review halfway through the program
• High level presentations from various business unit group leaders
• Weekly coffee chats with different ShopRunners (both tech and non-tech)
• Deep dives on applied machine learning from each data scientist on our team
• Meet and greet with the c-suite
• Regular demos of their work
• Final presentation
• ShopRunner outing event for fun

Intern Project Scoping

To define reasonable and meaningful project ideas, our team collaborated to identify opportunities. We created a poll to vote on projects that were achievable, clearly defined, had data available, and not time-sensitive. Through voting, our team surfaced five potential summer internship projects. We emailed the proposals to our interns and asked them to rank their top two choices. Doing so enabled our interns to work on projects they were interested in and passionate about. It also staked a sense of ownership in their project — even before it started.

Onboarding Document and Running Buddy

The first week in any tech organization is often chaotic. You are simultaneously setting up your coding environment, meeting the team, and learning about the business. It can truly feel overwhelming. On top of that, it may be your interns’ first time doing this. Onboarding documents help mitigate the stress while identifying specific steps that need to be accomplished during the first week.

We created an onboarding document which defined the schedule for the internship, mentors, role expectations, and laid out the purpose behind the various meetings already added to their calendars. Most importantly, we identified goals for the first week which included things like: logging into email, connecting in slack, installing docker, choosing a name for your github repo (another great way to stake ownership in an intern’s project) and creating it. Throughout the first week, mentorship effort was minimal as the interns knocked off each task in the document and came up to speed about ShopRunner.

In addition, we provided each intern with a ShopRunner running buddy. A running buddy is a colleague who works on a different team and can help answer general questions about the company, great places to eat, and generally help the intern feel more comfortable as they onboard.

Coffee Chats, Deep Dives, and Onboarding Talks

When the interns arrived the first week, their calendars were already filled with onboarding talks, coffee chats, and machine learning deep dives. Onboarding talks were led by the leaders of different business units (such as tech, devops, product, design) and were frontloaded in the first three weeks. They helped the interns understand the purposes and goals for each team and learn how ShopRunner works.

Coffee chats were set up weekly throughout the program. Initially, they were scheduled with data scientists so interns could better meet their teammates. By the middle of the program, they met with other engineers / designers / analysts to better understand how collaborations work.

Finally, we scheduled machine learning deep dives in the latter part of the internship. The purpose was to expose interns to the variety of work our team completes and broaden their knowledge about how data science is applied in industry.

Meet the C-Suite

As a startup, we reached out to our executives to ask if they would meet our interns during the program. Our CEO jumped on board and scheduled a thirty minute donut and coffee chat with them. Not only did our interns get an amazing opportunity to learn about the strategy of the business, but they were shown directly that the leaders of the business cared for their experiences at the company.

Demos

We wanted interns to participate in biweekly demos throughout their program. Once interns were generating results, our team helped coach them in demo-ing their work. By the end of the internship, the interns were presenting regularly. We also scheduled a final presentation in the last week, where we invited a larger audience including technology, data engineering, product, and executives. Interns were able to show off their work and continuing honing their presentation skills.

3.0 Management

Not only was it important for interns to develop skills and grow over the summer, but we also saw the opportunity for current employees to develop professional and management skills. Because many data scientists wanted to help support the program, we set up a structure where each intern had a primary and secondary mentor. The primary mentor would conduct 1:1s, project planning, and 360 reviews, while the secondary mentor would provide technical guidance or additional help when needed.

Project Planning and Direction

For each intern, we designed a project document which contained background reading, links to existing code, and a general progression of steps for the project to succeed. The first two months of the program were more explicitly defined while the last month was intended to be flexible, catering to the interests of each intern. We included interns work within our team’s agile kanban board for tracking progress over the summer. Initially, mentors helped plan out the two week agile sprints, but eventually allowed the interns to take over the planning process as they took on more ownership of their projects.

1:1s

To help our interns grow and develop, we chose to conduct weekly 1:1s throughout the program, and skip level 1:1s every other week with the director of data science. Skip levels provided a way for interns to receive less technical feedback and career advice, while also illuminating opportunities for a mentor to improve on managerial skills.

Because this may be your intern’s first 1:1 experience, it is important to have a plan and direction for each meeting. In preparation for these meetings, we reviewed tips on conducting 1:1s from our own managers as well as online resources. Some of the most useful links are below:

• Questions for our first 1:1
• 5 Steps to Building an Intern Feedback Loop for Program Success

During 1:1s, we focused on:

• Checking in
• Discussing situations that came up during the week
• Giving feedback (we use Spade and Diamond Feedback techniques)
• Asking for feedback

1:1s allowed us to modify project direction, encourage interns to take more ownership, and level set on where things stand. They also enabled mentors to develop critical communication and follow-up skills to ensure interns are able to accomplish their work.

Areas for growth

Interns develop quickly. They are eager, energetic, and want to make an impact. As a mentor, you have to delicately balance challenging them without it being too overwhelming or stressful. This is a tough act to get correct, and requires reading through the lines in 1:1s. The following areas were identified as opportunities for intern growth:

• Plan their work and manage their agile boards
• Review others’ pull requests
• Demo their work

360 Review

Feedback is one of the best tools a manager can use to help their employees develop and grow. At ShopRunner, we utilize a full 360 review process annually with a 6 month check-in halfway through the year. For our interns, we simplified our formal review process:

• Intern writes self review
• Mentor writes review
• Intern asks a colleague on our team to write review

Intern Self Assessment

1. Describe what has gone well thus far in the internship. Feel free to touch on your project, interactions with your teammates, and interactions with your mentor.
2. What can your mentor start doing or do differently to help you achieve your goals?
3. How would you most like to develop in the latter half of the internship?

Mentor Assessment

1. Describe what has gone well thus far in the internship. Feel free to touch on the project, and the interns interactions with teammates.
2. Describe opportunities for the intern’s development and growth

Colleague Assessment

1. Describe the intern’s strengths
2. Describe opportunities for the intern’s development and growth

Mentors then summarized the feedback and delivered to their respective interns. The whole process was completed in 2 weeks and provided additional insight for both interns and mentors.

4.0 Fun

Knowing this would be a transformative summer for the interns, it was critical that they enjoy their work and the people they work with. We designed a number of activities, sprinkled throughout the summer, to support community, relationships, and gratefulness we all experience working together at ShopRunner. Here are some events we planned for.

• First day lunch
• Game night
• Chicago Cubs baseball game
• Final day lunch
• Goodbye Taco-Bell themed party on our roof deck
• Goodbye gifts

Furthermore, we understood that both interns would be learning together as they developed code for different projects. Because of this, we sat our interns together so they could bond and teach each other tips and tricks they had picked up during the program. This led to our interns building a fun rapport with each other and allowed them to feel more at home within our team.

During the middle of our program, our company organized a Chicago Cubs Chicago Cubs baseball game which allowed employees and interns to get out of the office and enjoy a beautiful Chicago summer at Wrigley Field. This was a great way for all employees to leave behind work and enjoy time with their colleagues.

To celebrate the completion of the program, our team presented a gift of appreciation to each intern. After every member of our team signed it with a personal note, we presented the first Harry Potter book to each intern. These symbolically represented the naming scheme applied to the codebase our interns worked on. It also showed each intern how special the summer had been for us and that we were grateful for all the work they put in.

5.0 Retrospective

In the final stage, we created a review process to analyze what worked well and areas for improvement. We approached this via two routes: 1) we assessed from our own experiences organizing the program and 2) results from an HR exit interview process to anonymously report views of our interns. Here is what we learned.

What worked well

• Interns learned extensively through pull request reviews conducted by different members of our team.
• Great to have two interns who sat next to each other and were able to teach each other
• The two data science mentors shared the responsibility in planning. This ensured no detail was overlooked and the internship preparation stayed on track.
• Give interns ownership in their project. Let them choose it, name their repos, plan their work (with proper guidance), and own its success. It is much more rewarding for them and allows them to do their best work.
• Create a project and onboarding document to reduce anxiety in the first week and enable the interns to start strong.
• Show the interns how special they are. Schedule time with your C-suite and provide them the opportunity to share what they have done. Present them a meaningful gift at the end of the program.
• Skip level 1:1 between an intern and their mentor’s manager are critical for helping the mentor develop management skills.
• Present the interns with a gift at the end of the internship. It shows how much you appreciated their efforts over the summer.

Areas for improvement

• Coffee chats were great, but interns wanted more coffee chats with folks outside of data science. This is a great opportunity to learn about other parts of the organization.
• Desire for more intern-intern interaction.
• Setup more systems / logins / tools on laptop when interns arrive to make onboarding more seamless.
• Spreading out onboarding talks over the first month should be compressed into the first week. They were less relevant when interns were in the middle of the program.
• Interns wanted to learn more about tools / tips / tricks from different members of our team. This could be scheduled into a meeting at the beginning of the internship.

So, how did it go?

In the end, we were blessed to find two INCREDIBLE interns who both accepted full time job offers!!!

Best of luck in starting your next internship program!