Online technologies have made the flow of money easier, especially for illegal and criminal purposes. Financial Crimes, as the name implies, are not just restricted to monetary fraud alone but much more.
Paurav Pareek, our next pathbreaker, solves various problems related to Money Laundering, Terrorist Financing, Child Exploitation, Sanctions and Customer Compliance using Data Science and Statistical Modeling.
Paurav talks to Shyam Krishnamurthy from The Interview Portal about the most interesting part of his job being the real life challenges, which are extremely difficult to solve and the satisfaction of making a difference at the ground level.
For students, you can combine your love for numbers and statistics to uncover anomalies and iregular patterns, pointing to illegal activities, through Data Science and modeling.
Paurav, tell us about your background?
Well, I was born and brought up in Jaipur, Rajasthan.
My father was self-employed in commodity trading, so that created a particular interest in financials which led to a comfortable understanding of numbers and stats.
As I was above average in school I ended up in PCM stream which led to engineering (Computer Science). Though I was above average in school, as soon as I entered college, I had my ups and downs.
What did you do for graduation/post graduation?
I did my bachelors of technology in Computer Science from JECRC UDML Engineering College, part of JECRC foundation in Jaipur.
What made you choose such an offbeat, unconventional and unique career?
I still remember in the early days of my career I attended one boot camp program about machine learning. Although I cannot say I became any smarter after attending that boot camp the program certainly sparked an interest in me when the instructor said that “I can make predictions with 85% confidence.” It was like the fortune teller providing data as well behind his assumptions and predictions.
After the above session I developed a genuine interest in machine learning, especially its applications in real life.
Tell us about your career path
This journey started from my first year of college when I joined an ethical hacking club. I was hooked finding out different ideas and implementing them on the ground to know how technology worked behind the scenes.
This period turned out to be a very important juncture in my career, as it developed in me a curiosity to continuously improve on current knowledge without worrying about the fear of failing while trying something new. This approach has been benefiting me till date.
By the time I reached my fourth year, I had experience of writing algorithms and creating websites/projects by participating in different online competitions.
As I had a pretty good idea about moulding technological tools as per requirements, cracking a placement job interview was not tough.
Syntel Pvt. Ltd. (Now Atos-Syntel) was my first company. Most of my career progression took shape through “on the job” training, as I strongly believe in “on the job” training than any other means.
In the first six months of my career, I worked on various technologies like DB2, Shell scripting, VBA, Robotics process automation, Python etc., but was not satisfied.
That is when I attended the bootcamp on machine learning, which made me curious about this field. So I started exploring machine learning on google, searching for people online/offline who could guide me. I fortunately found some nice people who guided me initially.
The career transformation from technologist to a data guy took me one year and this was entirely guided by experienced folks with lots of theoretical learning followed by practical implementation on real time projects for big corporations during my tenure at Syntel.
MOOCs (Massive Open Online Courses) “Machine learning a-z by Kirill Eremenko”, “Data Science a-z by Kirill Eremenko”, “Machine learning by Andrew NG”, “Statistical Learning by Stanford university” and “Statistical Reasoning by Stanford university” helped me a lot in developing a basic understanding of Machine learning and the Data science field.
Next I developed a very strong desire to work in the field of data and my interest in numbers and statistics worked as a catalyst. So I took one more step further in my journey by reading books like: “Applied predictive Modelling by Kjell Johnson and Max Kuhn”, “The elements of statistical learning by Jerome H. Friedman, Robert Tibshirani, and Trevor Hastie”, “Think Stats: Probability and Statistics for Programmers by Allen B. Downey” and “Oreilly’s Hands on Machine-Learning With Scikit-Learn & Tensorflow”. These books became the building blocks for my journey and sites like Kaggle crafted my skills.
After accumulating some knowledge, I got the opportunity to work on my first data project. It was for one of the biggest financial institutes where i had to analyze the profitability of different regions on the basis of region specific environmental variables and to identify optimal parameters to generate maximum revenue per dollar invested. This was my first face to face experience with real life problems which were totally different from theoretical problems that i had encountered earlier. I had the chance of working with some of the most experienced people during this time. After this project, as a part of the same organization, I worked on a customer recommendation system.
One thing led to another and in some time I had a couple of several data science projects under my belt. Subsequently, I was offered work in a team in the area of Fraud Detection.
As soon as I was introduced to the team I knew I needed to be in this field, Financial Crimes.
How did you get your first break?
My first break was through campus placement.
What were the challenges? How did you address them?
As per my experience, If you have sorted out personal motivation and continue learning, then only one biggest hurdle will remain, which is to answer what do you want to make your career in?
Basically one broadly has two paths, 1) Go with the flow. Or 2) Try and choose.
Nothing is wrong in both choices above. Each has their own pros and cons, if one decides to go with the flow then this person has to be a quick learner, should have a general idea of things and how they work and should have a go getter attitude.
Latter choice is try and error approach (I am inclined to this)
Tell us about your current role
Currently I work for ANZ Bank, It’s one of the top financial institutes of Australia and New Zealand (Yupp!! the one on New Zealand cricket team’s jersey).
In ANZ, I’m part of the Financial Crime Threat Management Department (FCTM), which holds a very prestigious reputation in the Financial Crime domain among fellow institutes. I and my team fulfill the analytical needs of FCTM (Hence the name Financial Crime Analytics).
Most of the people think Financial Crimes as a victim-less crime as direct physical impact cannot be measured, but Financial Crimes range from simple credit card frauds, in-humane child exploitation, terrorist financing, human trafficking, sanctions compliance to money laundering. It is a $1.45 Trillion USD industry (Refinitiv Survey 2019) and Institutions are spending approximately $180.9 Billions USD(LexisNexis Survey 2020) to fight it.
In the analytics team we solve various problems which primarily fall within areas related to Money laundering, Sanctions, Terrorist Financing, Child exploitation and Customer compliance. We develop and implement money laundering detection scenarios using advanced analytics, solve customer segmentation problems, KYC problems and sanctions checks in real time and look back at the environment.
People in these teams should have good technical skills like: Data Analysis & Modeling, Statistical Analysis & Modelling, Rule & Risk based Modelling, Machine learning. They use Python, R, SAS, SQL, BigData, VBA, Power BI, QlikSense, Excel and PowerPoint on a day to day basis and can jump from one tool to another without any problem. But two abilities which are essential for everyone in these teams are Critical thinking and Problem Solving.
Working on different technologies in my initial years has helped me here and i am still learning and improving on some of the above abilities. This is a continuous process of improvement. If someone is interested in working in this field they can get technological experience using the MOOC course suggested earlier and try looking for similar kinds of profiles to get a real world exposure, which is always going to improve (No Upper bar here) your skills.
Typically on a day to day basis, we identify a problem, and then analyze and define it in terms of business impact. We then collect data around it, and try to solve it using multiple approaches (Excel to ML), get in touch with stakeholders for their insights, then again go back to collecting data, refining it and repeat the cycle till we reach an optimal and agreeable solution possible using the most suitable approach. We then present the results back to stakeholders with a future impact of the solution over multiple dimensions like (Operational workload, Risk acceptance, Compliance & Regulatory fulfillment etc.). After approval from various departments, we implement that solution and keep an eye on it for the next couple of months to monitor its performance and improve it if there are any shortcomings. We also monitor model drifting periodically.
The most interesting part of this job is real life challenges, which are extremely difficult to solve and one can get the satisfaction of real contribution by making a difference at the ground level.
How does your work benefit society?
In Financial Crime, our basis of any decision is society and its well being. We protect our customers from criminal entities, we work against child exploitation, and money laundering happening in various jurisdictions. We also curb terrorist funding. All the above activities have a direct impact on society and it provides self-satisfaction of contributing our bit to safeguard the community.
Tell us an example of a specific memorable work you did that is very close to you!
Well, I’ve several examples of work which is very close to me. One of them is from early memory when I was introduced to this field. I was working on a credit card fraud detection model for different jurisdictions by considering specific environmental variables. It was one of the most hectic undertakings as I was new to the Financial Crimes environment, so i had to understand different types of ISO payment messages, customer segmentation on the basis of their risk profile, different scenarios of clustering, Random Forest, Statistical modeling, Probability modeling and then combine all of them into a single risk rating. It was a wonderful experience and till date, that particular model is being used by organizations to safeguard customers against Credit Card Frauds.
Another one is in recent times. Sometime back, one the Financial institutes in Australia disclosed to the regulator that they had anomalies related to child exploitation activities in the system. It was a pretty significant failure from a compliance perspective in the Australian region. So we decided to look back at our data and create a model around it to check if we have missed anything? It was one of the biggest exercises we performed. We had to analyze various reports from different regulators to understand the problem, and then needed to create SQL models to run it over the last 10 years of transactions just to fetch data (extreme number of transactions were involved here). We performed one to many and many to one customer transactional mappings and a deep dive into each and every customer profile to understand their transaction pattern, and then developing suspicious pattern modeling through various rules and risks strategies.
Your advice to students based on your experience?
My advice is very simple. If you know your interest then one should work in that field only. It will be greatly beneficial in reward and satisfaction. For the rest of the folks like me, who are not sure about your career, try out different fields, different technologies, and work with various people. Always be in the company of more experienced/knowledgeable people than you and try to identify what you like doing.
Your work interests will continuously progress from one thing to another, but moulding a career which has a broader view and good understanding of internal working of things will be very beneficial in the long run.
Right now, I plan to continue doing what I’m doing and progress to different dimensions of the financial crime function (like: Trade Money Laundering). I want to enroll for a masters in business analytics but no timelines yet.