Let's get Technical
Here are some major projects I undertook in different roles.
01
Churn Reduction for Voot
As a Data Scientist, I addressed Voot's 35% subscriber churn issue by designing a ML model to segment users and predict at-risk customers. Tailored interventions led to a reduction in churn to 23% and an 18% revenue increase
02
Fraud Prevention for Paynearby Fintech
Tasked with fraud prevention, I leveraged ML to analyze anomalies within Paynearby's 1 million daily transactions. This resulted in a 92% cost reduction and a 4x boost in the fraud prevention team's productivity.
03
Optimization for Callarity Contact Center
At Callarity, I tackled inefficiencies and high attrition rates with data-driven strategies, including anomaly detection and personalized dashboards, resulting in a drop in attrition from 30% to 5% and a 35% revenue increase.
Other Projects
As a data scientist turned entrepreneur, I have had the opportunity to work on a variety of exciting projects. Here are some of the highlights of my career so far:
Project 01: Anomaly Detection System
Role: Data Scientist
Problem Statement: The challenge of detecting financial frauds and monitoring the platform for system failures to ensure the integrity and reliability of financial transactions.
Solution: Implemented a Deep Learning-based(Auto Encoders) anomaly detection system capable of identifying suspicious activities and system irregularities. This system leverages advanced analytics to monitor and flag potential frauds and failures.The output was presented as self serving PowerBI dashboard to risk team.
​Impact: Enhanced the security and reliability of the platform, also improved the understanding of user behaviour for risk and product team while reducing financial fraud incidents and ensuring a trustworthy environment for users and stakeholders.
Project 02: Customer Support Chatbot
Role: Data Scientist
Problem Statement: The need for an efficient solution to handle common user queries on the Voot Platform, aiming to improve customer support efficiency and user satisfaction.
Solution: Developed a chatbot using the RASA framework, tailored to resolve common user queries swiftly and accurately. The chatbot was designed with a focus on natural language understanding to provide a seamless user experience.
​Impact: Significantly improved the speed and quality of customer support responses, leading to reduced TAT by 95% and ~78% reduction in the workload of human customer support agents.
Project 03: Gamification for Productivity Enhancement
Role: Data Science and Strategy Lead
Problem Statement: Seeking innovative ways to boost employee productivity and engagement within the organization.
Solution: Implemented a system of gamification, introducing a leaderboard and smart incentives based on performance metrics. This system was designed to motivate employees through recognition and rewards.
​Impact: The gamification strategy led to a noticeable increase in productivity and morale, as employees became more engaged and motivated to excel in their roles.
Project 04: Credit Scoring from Alternate Sources of Data
Role: Data Science
Problem Statement: The need to enhance credit scoring methods using non-traditional data sources for a more comprehensive assessment of financial health.
Solution: Deployed a machine learning model that filters out financial red flags from SMS data using Natural Language Processing (NLP) and extracts financial information from text using Named Entity Recognition (NER). Additionally, the project involved customer profiling using transactional data to gain insights into customer behavior.
​Impact: This innovative approach to credit scoring allowed for a more nuanced and accurate assessment of potential borrowers' creditworthiness, leading to smarter lending decisions and reducing the risk of defaults.
Project 05: Data Warehousing with Google Cloud Platform
Role: Data Science and Strategy Lead
Problem Statement: The necessity to efficiently manage large volumes of data from production servers and SFTP buckets for better data accessibility and analysis.
Solution: Developed a robust data warehousing solution using Google Cloud Platform, which automates the process of reading data from various sources and writing it to Google BigQuery on a daily frequency. This solution streamlined data management processes.
​Impact: Improved data management efficiency, facilitating quicker and more reliable data analysis and decision-making across the organization.
Project 06: Analytics Infrastructure Development
Role: Data Science and Strategy Lead
Problem Statement: Seeking innovative ways to boost employee productivity and engagement within the organization.
Solution: We developed a robust analytics infrastructure to enhance our team’s productivity, which we then evolved into a SaaS product. This strategic transformation enabled us to shift from a service-based model to a product-centric SaaS model, opening up new revenue channels and market opportunities.
​Impact: Helped the startup pivot to a SaaS model from a service model.