Delivery Hero SE (DHSE, Berlin, Germany)
|
Data Science Manager
|
2023-2025
|
- Led Product Discovery and Personalisation team, focused on ranking and recommendation of products across both Restaurant and Q-commerce verticals of DH, increasing cross-sell conversion (~9%), average order value (~1.5%), and gross profit margins.
- Led the Incentive Science & Fraud team, focused on reducing incentive wastage by detecting fraud and optimising incentives across customers & restaurants, decreasing cost per user by 10%.
|
GoTo Financials (GoTo Group)
|
Senior Data Science Manager
|
2022
|
- Managed GoPay Data Science team for the user spend vertical, focused on improving user experience on GoJek and GoPay Apps.
- Implemented causal learning models to identify the most profitable customers/customer segments to run marketing campaigns, yielding optimal ROI.
|
MakeMyTrip India Pvt Ltd (MMYT)
|
Senior Principal Data Scientist
|
2016-2021
|
- Personalized Hotel Ranking & Discovery: Led the development of a common ranking framework and re-ranking models, increasing booking conversion by 30% and improving hotel recommendations with 14% higher conversion.
- Multi-Armed Bandits for Optimization: Implemented MAB techniques to optimize various business problems, improved decision-making through a data-driven, adaptive approach.
- Search Engine Marketing (SEM) Optimization: Built AI-driven bid optimization models using Genetic Algorithms, ACO, and ABC, improving search advertising efficiency.
- Cross-Platform Hotel Data Integration: Unified clickstream & booking data across platforms, using AWS Federated Query, creating an interactive dashboard for revenue, margin, and funnel analysis.
- Search Query Modeling (SQR): Mapped user search queries to relevant ad groups using Named Entity Recognition (NER), resulting in improved advertising relevance and targeting.
|
Fair Isaac India Software Pvt Ltd (FICO)
|
Lead Data Scientist
|
2009-2015
|
- Research, development, and delivery of predictive models for the retail and banking industry. Led the effort to develop analytic solutions into FICO Analytic Cloud. The models led to earnings of millions of dollars for clients across the US, Canada & Korea.
- Customer-level analysis for capturing attrition, lifetime value, and segmentation. These models were a part of the Analytic Offer Manager for the purpose of propensity prediction at the customer product level and demand forecasting.
- Enhanced Offer Prediction & Customer Profiling: Improved Analytic Offer Manager by integrating online browsing, GPS data, and multi-item offers, leading to better personalization and higher relevance in customer promotions.
- Recommender System & Market Basket Analysis: Developed collaborative filtering and LDA-based solutions for personalized recommendations; enhanced market basket analysis with modified Apriori algorithms, improving prediction precision & recall.
- Scalable Profile Generation: Built a portable Java-based profile generation supporting batch, incremental, and real-time modes.
- Demand Forecasting: Applied multiplicative time series models for retail demand forecasting, and evaluated the performance against ARIMA.
|