Why You Should Do a Research Internship and How to Get One

Rahim Shamsy
Atila Tech
Published in
10 min readMar 25, 2018

Source: Enjoy Teaching, 2015

What I did Last Summer

Last Summer I did a research project with the Electrical and Computer Engineering department on smart building energy consumption predictions. You can see some of my work here. This is a field where I had limited exposure to and no relationships with professors doing research in this field. Despite those initial shortcomings I was able to get a research position where I learned a lot. This article will explain how I was able to get a research internship despite my limited initial knowledge, some advice on how you can do the same, and a few important lessons I learned.

Rahim is in the civil engineering and business program at the Ivey Business School in Canada. After graduation, he plans to learn more about machine learning and blockchain technologies and apply those skills to help the various industries in his home country of Tanzania.

Originally published at the Atila Blog.

Why do Research?

In its most basic sense, research is the process of investigating how a problem can be solved. The ‘scientific’ process — the constant hypothesizing (guessing what the solution is), creating tests (to prove those guesses right or wrong), testing hypotheses, gaining insight, hypothesizing again, and repeating the process — is a great tool in figuring that out. Having a knack for experimenting is truly valuable, not only in research but in one’s career. Research therefor prepares you for a way of thinking about problems. It keeps you humble — you don’t know the answer to a question, but you have calculated guesses of what they may be, so you test it out. This process is the backbone of anything that requires creation, be it a startup (check out The Lean Startup by Eric Ries), a product (that requires design thinking), or creating knowledge for the benefit of others. Research puts you in the habit of cycling through this scientific method, and that is invaluable.

Academic research, as far as my experiences go, requires learning quickly, and applying those learnings as you go. As an example, before I could begin working with the building’s energy data, I had to learn about statistical modelling tools like machine learning, and what form of data was necessary. I then had to translate this into visualizing, cleaning and wrangling the data, using Python packages such as Matplotlib, seaborn, Pandas and SciPy to work with the data. Learning quickly and applying learnings effectively was very important, and this is a skill I have developed through research, thankfully. Not only that, but getting research guidance and supervision from professors and post-doctoral researchers was immensely helpful, as I could draw from others’ learnings and adopt best practices. Best practices, and doing what you are doing to the best standard is very pivotal in research and in one’s career in general.

How is Research organized, and what types of research are there?

Professors at universities are usually part of broader research groups, which share a research theme. Each professor would specialize in a certain sub-area that they are interested in; they will have multiple students they supervise in research projects. You should browse the ‘Research’ link on your university/faculty-of-interest’s website, and browse through the various research work that is performed. As an example, Civil Engineering research work done at my school can be found in this link — the links in this website lead to specific research groups (e.g. FIDS).

Every research project undertaken by professors at universities are aimed at solving certain problems. Nonetheless, some problems that professors are working on involve directly solving a practical issue (e.g. projecting insurance risks based on climate modelling, using statistics), whereas others involve indirectly solving practical issues (e.g. making improvements to statistical modelling tools, which may then be used to more accurately project insurance risk). Making this distinction when browsing research work is important, and I advise you to understand what type of research you want to work on. This will help you narrow down the various opportunities. Both types have their benefits and drawbacks, but it all depends on what YOU are after, and you should pursue opportunities YOU are most interested in.

How I Got the Research Position

Machine learning is a field that I’m very interested in and my goal is to start a company in that field. But, I felt like there was a skill gap between what I knew and what I needed to know in order to build something. I felt like a research internship would be a great way to learn a lot about a very technical field while being advised by knowledgeable professors and fellow researchers, and gaining industry exposure.

Once I knew that I wanted to do a research internship, I sifted through researchers’ profiles at Western, and came across my supervisor’s research group. I also reached out to one of my professors (who taught me in first year), expressing my interest in learning about applying machine learning — I learned about research work being done in the area at Western. Doing both (researchers’ websites, and reaching out to professors) exposed me to multiple options — I used this information to cast a wide net.

I was particularly interested in a professor, who applied machine learning to smart building technologies within the ECE department at Western University. As a Civil Engineering student, I had come across the statistic that buildings consume ~40% of electricity (in America at least) — this to me was an area where energy efficiency could lead to positive environmental effects. How could I resist such an opportunity ? Learn how to apply something I was interested in, and create a positive impact!

Usually, professors link information for summer undergraduate researchers on their personal or research-group websites. That would be the first place to start — shows who wants help for the summer, and who may not. Instead of blindly applying, I would suggest you go speak to the professor, especially if you lack the skills. But before you go meet the professor, you need to show that what you lack in skills, you make up for in hard work and commitment to continuous learning. Learn something related in the field and SHOW that you are committed, rather than just saying that you are. This is essentially what I did to secure the opportunity.

The first time I emailed my professor, she made it clear that she was very busy and that she couldn’t meet. At that point, I thought to keep the conversation going, and mentioned how I would like to meet for advice on how to learn in order to build skills for research in the future. Once she returned from her travels, she agreed to meet me and from there it was how well I sold my learning abilities. I truly had to hustle to learn and get to meet her. Lots of luck was involved, but hard work as well.

Seeing that I was dedicated to learn and contribute to the group’s work, she brought me on as a Summer Research Associate and I was fortunate to get a scholarship as well. I was offered a position working with the university’s Facilities Management department on on-campus building efficiency.

What You Should Do

In summary, the steps that I think YOU should follow if you want to seek a position:

  1. Think about where you want to be in a few years, and work backwards to figure out what skills you need (Directly vs Indirectly Applied Research; Which domain-related problem you want to explore (pharmaceuticals, materials science, clean tech, gaming, manufacturing, etc.); What researching tools do you want to learn)
  2. Reach out to professors who you have a good relationship with (you should seek to build these relationships throughout undergrad — they are very knowledgeable and truly want to help). Ask them about what research they know of, where you can build those skills and apply them
  3. Do your own research — your faculty/faculty-of-interest will have a ‘Research’ link on their website. Look at the various research groups, and research professors. Start by a first pass — shortlist those that interest you while you read about their work.
  4. Learn more about the work that each of the shortlisted professors do. Do your own research in those areas (Google!), and make sure they: a) have a strong link to your skill development goals; b) Pique your interest.
  5. You should arrive at 3–5 professors you would really want to work with.
  6. Start with your top-choice, and do some self-learning if you don’t know anything about the field they are in/the tools they use. I didn’t know much about Machine Learning, so I went ahead and did a mini online course.
  7. Reach out to the professor, mention how you are interested in the work they do, tell them you are interested in the area they are doing research in, and the tools they are using. Relate to what you have learned about the tools, and ask them specific questions about how they are using those tools in their research. The goal here is to be genuinely curious, so ask them anything about their work that makes you curious and that relates to what you have learned. Arrange a meeting with them, but make it about learning from them (Not about you doing summer research with them!).
  8. When you meet the professor, speak about their work, and ask those specific questions. Start with general questions (why they research in that area, and what work they have done in the past) and then go to specific (asking questions about the specific research that interests you). Most of the conversation should be ABOUT the research work they do (details). Mention how you have done self-learning, and that is what makes you curious about the research area. Ideally, they should see your interest as commendable, and that makes it easier for the next step.
  9. Ask them in the meeting — towards the end — if you can help with the research they are doing because (as you will have proven) — it genuinely piques your interest.

Other Thoughts on Research:

Domain Expertise and Problem Finding

Sometimes, research in one area will make you realize of other areas that require problem-solving. This could manifest itself as a business idea, or a further research project idea. Over time, I have noticed that successful companies have been created as a result of commercializing academic research. More specifically by applying researched problems to practical scenarios. There are many big (and obvious) companies in this categories, such as Google, but there are many less-mainstream companies such as Maluuba (acquired by Microsoft), ModiFace (acquired by L’Oreal), that have been started either through specific-research or by experienced researchers.

How to get paid to learn

There is lots of money is available for research in the form of scholarships and grants. Here are some good resources to get you started:

What I Would Do Differently

From the onset, my research project was loosely defined. This could be a good thing, as it may allow for the flexibility that IS required in research, as you are constantly iterating and updating your assumptions. You MUST be comfortable with uncertainty if you are to do research — from the definition of the research project, to how you would answer the questions asked. Perhaps having a loosely defined project is not best practice; perhaps ‘best practice’ is relative. Regardless, I wish I ‘embraced the uncertainty’ more than I did. You need to realize that you are getting frustrated, and you need to stop. I thank my post-doctoral advisors who consistently reminded me of this, and that helped me a lot. After all, I was getting a real taste of research — not too sure how I feel about that still.

Because my supervisor was on sabbatical leave during the period (and therefore intermittently on campus), it was difficult to build a working relationship with her. Nonetheless, the post-doctoral researchers in her group made up for her absence, for which I am truly grateful. This was surprising to me in the beginning, as I was expecting to work closely with the head of the group (the professor) herself. As far as expectations go, you should make sure who you will work with, as that may affect whether or not the learning opportunity is attractive — what you will achieve may be a function of how strong your direct-supervisors are. I was lucky to have worked with post-docs and other professors who were truly helpful, but it would have helped to have that clear from the onset.

Conclusion

There’s lots that I could talk about, but the most important things are the following:

  1. Self-learning is an exceptional skill to have, and I have not witnessed faster learners (and ‘unlearners’) than the post-docs at the research group I was with. They were notorious about learning quickly, learning the most important aspects of what you need, and quickly applying what you learn to create something valuable. I am nowhere near emulating the level of learning these postdocs were undertaking, but I am much closer than I was before. It has helped me to continue learn about a variety of topics I have been interested in over the past year — Machine Learning, how the Internet works, how Blockchain works, and many more.
  2. Research is a great way to bridge your schooling knowledge to real-life application of the knowledge. This will prepare you for more fruitful career opportunities, as it will equip you with invaluable skills and knowledge specific to the domain you are researching in, and general to research.
  3. In the end, as hard as you may work, there may be factors out of your control. For me, the quality of the data I was provided was the real constraint, as it lacked the specificity and feature-detailing required for creating an accurate supervised predictive model. Coming to terms with this possibility may prepare you better for such situations. Again, as a first-time undergraduate researcher, it should be about learning and the process of research — not the outcome. With subsequent research projects, focusing on the result may be more wise.
  4. You truly need to show the professor whose work you are interested in, that you bring something to the table and are coachable. Doing some learning on your own will give you a taste of their work, and will allow you to

If there is anything I have missed, or gotten incorrect, I would love to know so I can learn about it. Ask me a question if you have any — reach out to me on Twitter @RahimShamsy.

Thank you Tomiwa for your help on this, and for showing me how to learn by sharing.

--

--

Rahim Shamsy
Atila Tech

Exploring tools to make people explore interests