My experience with data and machine learning began when I started the applied data science master’s program at the iSchool. I formed a foundation in ML via classes such as Introduction to Data Science, Applied Machine Learning, and Quantitative Reasoning for Data Science. This provided a foundation of everything I worked on during the internship. In fact, my project work for Applied Machine Learning was directly linked to my landing the internship, to begin with!
My Background
Computers and programming come naturally to many of us, as do the rigors of mathematics and statistics. For a lot of us, this inclination towards the wonderful world of coding and numbers has been a crucial part of our lives – having studied said subjects from an early age, a lot of us continue the practice and eventually find ourselves in well-paid and respected positions as computer engineers, data scientists, data engineers, et al.
I was not one of these few. My tryst with computer science and mathematics ended at the age of 15 – or so I thought. My undergraduate degree was in International Relations and International Business, and post this, I worked in Business Development.
I tell you this to establish my lack of technical credentials (beyond an earnest curiosity and some initial elementary exploration of the world of data science) before enrolling in the MS-Applied Data Science course at the iSchool, Syracuse University, and to reinforce the fact that one does not need a storied past working with computers and data, or incredible knowledge of complex algorithms to land a rewarding and technical internship.
In reflecting on my internship experience, here are some of the most important take-aways from the internship application process, interviews, and my ten weeks working as an Intelligent Process Automation Engineer at Chubb.
Applications
Working with the Career Services at the iSchool early ensured that I had my resumé and CV ready. Further, speaking with their counselors helped me understand how to approach the job hunt; from identifying my skill set and desire to add technical knowledge to my portfolio to ideating a plan of action to approach recruiters and companies; speaking with said counselors helped create the platform upon which my internship hunt was predicated.
An important point to remember here is that rejection is a part of the process; changing your mindset towards rejected applications from one of woe or despair to that of learning – of understanding why you may have fallen short and how you can improve is vital. Career Services was incredibly supportive here, and their counselors were always on-hand to ensure I never felt too poorly about application rejections.
Interviews
Depending on the firm and position you apply to, you may have Code tests, Online Assessments, and live interview rounds. The specifics of the process are mutable and differ.
My interview experience lasted two rounds with no online assessment or live coding. My initial round was a forthright conversation with HR. This was followed by a longer (1.5 or so hours) round that saw me interact with my team manager and the CIO for Intelligent Process Automation at Chubb.
The interview had me walk through my final project for Applied Machine Learning and explain the ML algorithms used and why those algorithms, in specific, were used. I was asked about the assumptions different algorithms make; there were also a series of questions designed to help my interviewers understand how I solve problems, how I interact with teams, and how I might work in the techno-cultural environment at Chubb.
Confidence and ease of communication were critical to my interviews; I knew my resume inside out, I knew my project work and the research that underpinned things I had done, and most importantly, I was happy to communicate my ideas through conversation and not through monologue; fostering a discussion with my interviewers has helped me establish a professional relationship of advice and mentorship that continues even now though my internship has finished.
The Internship at Chubb
My internship saw me work as an Intelligent Process Automation Engineer at Chubb, the world’s largest P&C insurer; the internship lasted ten weeks – from the 2nd week of June until the end of August and had me work out of Chubb’s Jersey City office.
The first two weeks at Chubb focused on using their WorkFusion software and getting a plethora of certifications. During this time, I sat through countless hours of tutorials and assignments focused on certifications for machine learning, data science, and WorkFusion’s platform. In addition, the first few weeks also saw me busy with a bevy of activities focused on helping students assimilate into the culture at Chubb and getting to know the team I would be working with.
From Week 3 of my internship onwards, I was assigned to work on an email classification and routing automation project. In brief, we took business rules from the subject matter experts in the business operations team and used said rules to classify emails. These classified emails were then sent to the appropriate team in the proper insurance stage (agency, brokerage, legal, claims, etc.). Initially, I used Python to create a simple rule-based classifier; reportage based on this classifier was assembled and presented as Excel spreadsheets. Over time, we realized that a minimal subset of mails (<50%) could be effectively classified based on said rules.
This realization inspired my internship project, on which I worked from Week 5 to the end. Having realized that business rules were difficult to grasp while not being as effective as hoped, I worked with a fellow intern to develop an ML-based classification system; however, we had to implement code for data identification, cleaning, processing, and storage.
Since I was working with real-world emails, and since a number of these emails had attachments, I had first to extract the email text and identify whether the mail had attachments and the kind of attachments said mail had. When the attachment was a pdf, I also had to understand if the pdf was searchable, and if not, then implement OCR (we used PyTesseract for this). Essentially, this internship had me work with unstructured data, which I then endeavored to clean and process. Once the data was in a usable state (text only), I ran it through numerous algorithms before using a Random Forest based classifier. This, in turn, increased the efficiency of email classification from the high 40s to the high 80s!’
Reflecting on my Internship
Working at Chubb, beyond giving me confidence in my ability to work with computers and data and to write efficient code to contend with the issues we come across in data, also helped me acclimate to the professional environment in the USA. I learned to deal with problems not only by trying to find solutions myself but through collaborative communication; most importantly, working at Chubb allowed me to take lessons learned in classes like Business Analytics, Data Analysis, and Decision Making, Applied Machine Learning, and Quantitative Reasoning for Data Science and use them to solve real-world problems.
In fact, my Applied Machine Learning class gave me most of the tools I would need to work with data at Chubb! Everything was dealt with in class, from data processing and cleaning, transformations, and PCA, to applying ML algorithms and selecting the best model. Further Business Analytics and Data Analysis and Decision Making both helped me work with Excel – this also saw me work with teams that were not my own just to show them how they could optimize their Excel use using tools like Pivot Tables, for instance. Finally, Quantitative Reasoning for Data Science was integral to my ability to make sense of statistics, make models fit, and ensure that the data was presented in a format our algorithm could work with.
Hints on Finding an Internship
- Have your resume, CV, and Cover Letter ready to work with as soon as possible. Meet Career Services early and meet them often.
- Identify fields of interest and seek internships within those fields, focus on what you can learn and how the internship in question can/will help you gain the insight and experience you seek
- Rejections are a part of the process. Accept them as learning experiences; use the feedback you get from every rejection to grow as a student and as a potential professional.
- Apply early, apply often, apply a lot, and focus on your applications. You have very little to lose by applying and a lot to lose by not applying.
- Tailor your resume to the specific role you are applying to; not all positions have the same requirements; ensure that your resume answers the needs of the job where you can.
- Be confident in what you know but accept when you do not know something. Accepting your limitations and identifying how you can work through those limitations is better for you than being woefully wrong.
- Review your resume, and ensure you know what is on the said document before your interview – this is not entirely necessary. Still, I’ve always found it an easy way to boost confidence in an interview.
- Allow for a conversational environment – while an interview is a formal setting, there is no reason for it to be uncomfortable for any of the parties involved. Treating it as a technical conversation takes much of the pressure off your shoulders, making the interview much easier to get through.
- Networking is vital; networks won’t just help you secure professional roles. People in your network can be founts of information to help you quickly solve seemingly complex problems. Your manager, colleagues, and the company you are interning with want you to grow – they want you to succeed. Remember that and use the resources on hand to grow.
- Be confident in what you know, and back yourself to learn what you don’t know. One of the most astounding human skills is our ability to learn not just from ourselves but from others – use the internet, speak to your bosses, interact with your colleagues, and take every opportunity you get to better yourself.
- Take time out. It is terribly important to learn and to grow, but you also need time to rest and recuperate – don’t skimp out on it.
Have fun. You always learn more when you can have fun – participate in workplace events. Chubb, for instance, had a lovely tech expo and regular climbing sessions in the city. Find events you enjoy wherever you are and participate in them!
