This week we bring you a closer look at Kabbage, a Fintech company headquartered at Atlanta USA, through an interaction with Ratnakar Pandey. RP, as he is popularly known, heads the Analytics and Data Science portfolio at Kabbage India.
Kabbage has provided over $3 Billion line of credit loans to over 100,000 businesses. Kabbage is one of the fintech unicorns and has achieved various accolades such as 36th fastest growing company in the US on the INC 500 List, Fast Company’s, Top 10 most innovative companies in finance, Forbes, America’s Top 100 Most Promising Companies etc. Kabbage is more than a lender for small businesses; our data and technology platform is now being used as a fully branded product by other lenders, and our products are expanding.
RP has more than 15 years of experience in analytics and data science fields. At Kabbage, RP is leading the machine and deep learning models development activity across customer lifecycle, from acquisition to customer engagement to fraud prevention to risk based underwriting policy development.
Prior to joining Kabbage, he has been a part of the data science leadership team in Citigroup, Target, Texas Instruments, and few startups in India. An MBA from ISB Hyderabad and MS from University of Arkansas, below are the interview excerpts of Ratnakar Pandey with Analytics India Magazine, where he talks about analytics at Kabbage, skills they look while hiring data scientists, the working culture at Kabbage and much more.
Analytics India Magazine- Could you tell us more about Kabbage and your data and analytics platform?
Ratnakar Pandey- Kabbage significantly improves upon the conventional lending model that are championed by banks and other similar lenders. Most of the heavy lifting that happens in Kabbage is done by technology and data, which is one of the key differentiators. We like to call ourselves a data company rather than a typical finance company. So, data is the core competency of the company and we use that in our decision making every day.
We have 350+ employees worldwide, and I manage a team of roughly 45 data science professionals in India. The work that we do is centered around the US market currently, however team is increasingly getting involved in providing data and analytics support to Kabbage’ s partners with global footprint such as ING and Scotiabank
AIM- What benefits have companies reaped using the analytics solutions offered by you? Would you like to highlight few use cases?
RP- As we discussed earlier, data and technology is at the core of all decision making in Kabbage. There are several great benefits that the company is deriving from the analytics and data science. Let me highlight couple of examples-
First, data science and technology enables us to make fast, automated and accurate underwriting decisions leveraging our proprietary machine learning models. Customer fills out a simple loan application online or on our app, Kabbage’s system take care of rest of things. Kabbage connects to a dozen plus diverse data sources such as Yodlee, Paypal, QuickBooks, Square and pulls the info such as business revenues, cash flow, ratings etc. to ascertain businesses creditworthiness and to determine loan terms and conditions. The most interesting thing is that all this happens within few mins in a completely automated fashion with no paper documentation. Because of this, a large majority of Kabbage’ s customers have a totally automated, digital and seamless loan application experience and money get into their accounts without any hassles.
Another notable example of data science usage is in driving existing customer management strategies to drive maximum customer satisfaction and high customer lifetime value for Kabbage. We do it by following a multipronged approach such as which industries to target, time and channels for contacting customer, proactive credit line management, design of creative, collection strategies etc.
AIM- Would you like to share some of the analytics solutions you have worked on?
RP- We use several programming and data management tools for providing both tactical and strategic analytical solutions, some noteworthy tools are Python, R, SQL, Spark, Hadoop and Scala. In terms of statistical techniques for machine learning, we routinely use regression techniques (linear, logistics) and classifiers such as Gradient Boosting Machines (GBM), Elastic Net Regression, Support Vector Machine (SVM), Ensemble Learning etc. For forecasting and anomaly detection we use ARIMA/X, k-NN and other similar techniques. We also heavily use Natural Language Processing (NLP) for drawing insights from text and unstructured data.
AIM- What is the roadmap/ plans for analytics at your company in the future?
RP- In lending business it is imperative that we do a great job in risk and fraud analytics to safeguard company’s and stakeholders interest. The excellence that we have shown in these areas will continue to be the case in future as well. At the same time, we are building deeper expertise in marketing analytics, particularly portfolio segmentation, campaigns management, A/B testing among others.
Furthermore, since we are routinely dealing with thousands of features and variety of data sources therefore we are getting more entrenched with deep learning techniques such RNN, CNN, LSTM, Autoencoders using Tensorflow and Keras.
AIM- What kind of knowledge and skill-sets do you look for, while recruiting your workforce?
RP- For the positions that we are currently hiring, we are looking for 8+ years of work experience in data science, preferably in the banking and lending industry. We are keen on hiring people who can lead a project from start to finish and take the full accountability and ownership of it. We generally hire from the premiere institutes, 90% of people we have right now are from tier 1 institutes. Moving forward, we are looking forward to have a more diverse workforce.
We have also started hiring interns, which is something we didn’t do earlier. We hired couple of interns from the Business Analytics program from a college in Pune. We are also in discussion with Aegis, Mumbai, which has a collaboration with IBM.
In terms of skills, we look for Python, Spark, SQL, Scala, R, familiarity with big data, KAFKA environment and related technologies. Important to reiterate here that we were featured in Glassdoor’s 2017 Best Places to Work.
AIM- What are the most significant challenges you face being at the forefront in analytics space?
RP- Lack of machine learning and deep learning talent is the biggest challenge in the Indian market. It takes at least a couple of months or longer to hire the right candidate. Even if you find the right candidate, offer decline and dropout rate is quite significant for the best talent available in this space. To complicate the picture, there is a huge mismatch in the demand and supply. Demand is very high and supply of the quality talent is very less.
The other challenge is how do we quickly learn the latest in the industry. Technologies are evolving very fast and you must make sure that you invest significant amount of time in learning and development on a proactive basis to stay ahead in the frontier.
AIM- How do you think ‘Analytics’ as an industry is evolving today? Could you tell us the most important contemporary trends that you see emerging in the present analytics space across the globe?
RP- Present-day customer wants to have a right product or service at the right time and in the right place. We are generating more data than ever before- 90% of the data that we have today is generated in last 2 years alone. This data is coming from a variety of different sources such as voice, text, transaction, sensor, chat, images, videos etc. To handle this fast moving, heterogenous and multimodal data we need to get more entrenched with machine learning and deep learning to make real time analytics driven decisions that will bring maximum value for the customers and companies alike.
Subscribe to our NewsletterGet the latest updates and relevant offers by sharing your email.
Srishti currently works as Associate Editor at Analytics India Magazine. When not covering the analytics news, editing and writing articles, she could be found reading or capturing thoughts into pictures.