We thank Siddharth Uppal from Citibank N.A. for taking part in this interview and sharing his experience as a Data Scientist in the Banking & Finance domain. Siddharth being from a Mechanical Engineering background found his way into the finance domain and soon realized the importance of Data Science & AI in the Fintech industry.
In this interview, Siddharth talks about his way of doing Data Science and some tips to do well in the field. He also talks about how to crack the interviews of Fintech companies for Data Science roles. He also shared some insights from the industry like:
- How to prepare for coding for interviews
- Which books to read for ML
- Advancements of Finance domain into ML
AI/ML in the banking and finance sector provides with faster, more accurate credit decisions and insights about transactions on multiple credit products.-Siddharth Uppal
Start your day early and practice coding before you begin your workday. This reinvigorates the mind to think analytically and strategically to contribute effectively to a data science team.-Siddharth Uppal
Let’s hear from Siddharth now!
From a Mechanical Engineer to a Data Scientist
You hold a bachelor’s degree in Mechanical Engineering. How can a Mechanical Engineer become a Data Scientist, and what made you decide to pursue a career in this field?
Mechanical Engineering coursework is generally focused on computational and advanced mathematics with some exposure to coding on MATLAB and Stata. My coursework in Mechanical Engineering was no different and during my study, I was equipped with the theoretical knowledge of statistics and probability in great detail. In data science, a data scientist simply employs coding tools and analytical software to realize the practical implications of statistics and probability which benefits the senior stakeholders to make concrete decisions with some evidence that we delineate from data.
During my study in my undergraduate, I always wanted to pursue a career in finance with a special focus on consumer products in a leading financial institution. After working in the traditional construction/automotive industry, I was sure about my career option in finance as data science was making significant strides in streamlining and transforming the industry in contrast to the traditional construction/automotive industry. This career path was complemented with a master’s degree in engineering with extensive coursework on data science and analytics.
AI/ML upskilling at Citi
How is Citi encouraging its current employees to upskill in AI/ML? What steps and measures are being taken to grow AI teams in-house?
Citi fosters a culture of collaboration and partnership for every employee irrespective of their department and division. This enables anyone interested in data science to concentrate and develop specific coding skills with direct mentorship from our seniors and the leadership team. Citi also has seminars on multiple topics such as data science, machine learning, etc. to keep all employees abreast with the newer technologies and software which are evolving in real-time.
Besides these seminars, Citi also offers online courses with multiple e-learning websites to get a crash course on coding and theoretical knowledge related to data science. In my team, for every new hire, seniors and peers take significant time to introduce them to the structure of data and how data science makes an impact in our overall business ensuring that they have a successful career in data science and significant contribution to the company. The onboarding process is similar to a peer to peer, teaching model, where a mentee is introduced about data science and practical applications of machine learning in the initial months before they can dive into real-world projects and assignments.
Coding preparation for interviews
As a data science leader in a global investment bank, can you share some tips on how to prepare for coding rounds for financial institutions?
Coding rounds are very diverse and depend on the specific of the role applied in multiple companies. To prepare well for coding rounds covering the majority of the roles in the market, one needs to start with the basics of R, Python, SQL and statistics that are regularly tested to see the fit and the competency of the interviewee. I relied on available learning sources such as Hackerrank, Leetcode and Datacamp to practice multiple questions that can be a part of the verbal and/or live – coding rounds of the interview.
I also recommend ‘Machine Learning with Python Cookbook’ by Chris Albon which gives practical solutions from preprocessing to deep learning in data science. There are many free videos readily available on YouTube as well by professional data scientists that cover important interview questions with a walkthrough approach on how one should best answer the interview questions in data science roles.
Applications of AI/ML in Finance
Where do you see the biggest areas of improvement for AI/ML in the Banking and Finance sector? How do you think AI is handling the criticality of this industry applications such as Fraud monitoring, Anti Money Laundering, etc.?
AI/ML in the banking and finance sector provides with faster, more accurate credit decisions and insights about transactions on multiple credit products. ML has enabled to accurately flag and blocks potential fraudsters preventing fraud even before fraud occurs on the system. This enables a higher degree of customer satisfaction and prevents significant fraud losses improving fraud monitoring in real-time.
AI/ML also improves a customer’s experience on digital products which has seen significant growth in the past few months during COVID – 19. Fraudsters always find new ways to commit fraud on multiple consumer products, ML gives an insight into specific areas that need focus to strengthen fraud detection and prevention strategies.
Skills required for being a Data Scientist
What skills and attitudes do you look for when hiring data scientists?
For my current role at Citibank – Global Fraud Prevention team, I am responsible for managing fraud prevention strategies on digital channels for Credit Cards, Retail Services, and Citi Private Bank to reduce fraud exposure, and improve identification of fraud patterns; minimizing fraud loss using R, SAS, SQL, and Machine Learning techniques.
For hiring data scientists, the team generally looks for three distinctive traits that are necessary to be successful:
- Curiosity to learn and grow
- Communication with senior stakeholders and partners
- Coding competency on multiple software and analytical tools
Day to day work @ Citi
What is the day to day responsibilities of a Data Scientist at Citi?
I begin my day by analyzing digital customer data to see patterns in fraud events and identify how fraudsters try to compromise our system. This is followed by a series of meetings (or Zoom meetings during Covid – 19) with multiple stakeholders presenting my findings and proposing solutions/recommendations to strengthen our fraud detection system. Once everyone is on board with the recommendations and analytical solutions, we make changes to our fraud strategies to prevent future fraud events. This job involves extensive use of data visualization tools such as Tableau and Power BI to develop and share insights related to the fraud trends on our consumer products (Retail and Credit Cards) to multifarious teams such as Risk, Product and Legal as well as to the US Consumer banking leadership.
My analytical role has resulted in the prevention of significant fraud losses and rapid identification of fraud events with non-monetary and monetary transactions by analyzing customer behaviour for improving authentication strategies and to accurately detect fraud on compromised accounts.
What unusual or absurd thing do you practice or advocate for in your role as a data science leader?
Start your day early and practice coding before you begin your workday. This reinvigorates the mind to think analytically and strategically to contribute effectively to a data science team.
What is one book that you would recommend young data scientists to read?