Abhijeet Bali

• Boosted company revenue by impressive 25% through smart data-driven strategies. Created advanced machine learning models and personalised recommendation systems that played key role in increasing profits.
• Took lead in automating important projects related to understanding customers, improving business processes, and enhancing product insights.

Key Projects:

Lookalike Modelling:

We implemented Parallel profile Modelling for audience extension, utilizing advanced techniques such as K means++, silhouette score, PCA, and FAISS. This intelligent approach ensured precision in identifying similar customers within universe dataset, avoiding audience saturation. By employing these smart methods, we not only expanded our audience but also enhanced the efficiency of our outreach strategies, enabling the discovery of more customers with shared characteristics and interests.

Category Retention Model:
Developed an advanced Category Affinity Model, evaluating the likelihood of customer purchases across diverse categories for a massive one billion customers. Leveraging tools like XG Boost, statistical tests, and comprehensive data analysis, the implementation significantly enhanced the efficacy of marketing campaigns, resulting in substantial improvements in revenue generation.

Segmentation Engine :

Contributed to the development of a segmentation engine product by creating distinct segments and personas based on customer data analysis.

This allowed seamless execution of marketing campaigns and precise mapping of offers to customers, enhancing reach and promoting specific product categories effectively.

Data Engineering Pipeline Development and Monitoring:

• Development and creation of robust data engineering pipelines, tailoring them for diverse use cases and ensuring seamless data flow into our database.
• Established comprehensive Data Quality Management (DQM) system to uphold data integrity.
• Implemented scheduling through Apache Airflow, ensuring reliable and timely execution of all pipelines for effective database management and optimized data flow

Reporting and Dashboarding :

• Partner Health Dashboard: Partner Health Report Power BI Dashboard provides a concise overview of partner spending, redemption opportunities, and key metrics such as ATV (Average Transaction Value), ACV (Annual Contract Value), ATF (Annualized Transaction Frequency), and customer distribution between new, existing, reactivated, and churned. Dashboard also presents revenue insights, including current run rate, target, and required run rate, tracking performance throughout fiscal year till last month.
• Redemption Opportunity Loss (ROL) Dashboard: Redemption Opportunity Loss Power BI Dashboard presents a dynamic analysis of opportunity loss trends month-over-month. It provides insights into store-wise opportunity loss and Point Bucket-wise Return (ROL) relative to invoice amounts. Additionally, dashboard showcases available point buckets concerning non-redeemed transactions, offering a comprehensive view of potential missed opportunities and contributing factors.
• Adhoc Data Request: Collaborated closely with partner managers and partnership team to fulfil data requests aimed at evaluating the efficacy of sales campaigns. Utilized analytical skills to extract and analyze relevant data, subsequently generating comprehensive reports showcasing campaign performance metrics.

Implemented a cutting-edge fraud detection system, reducing fraudulent transactions by 30%. Proven expertise in data analysis, feature engineering, and cross-functional collaboration.

Key Projects :

Risk Monitoring for Consumer Finance:
Conducting fraud analysis for customers approved for private loans through machine learning involves employing advanced algorithms to identify unusual patterns, behaviors, or anomalies that could be indicative of fraudulent activity.

Transaction Pattern Analysis:
The predictive model anticipates future transactions within next three months based on historical customer behaviour. Through application of advanced bagging and boosting techniques, model identifies most promising customers for targeted campaigns. This strategic approach directly enhances revenue by concentrating efforts on engaging customers with highest probability of conversion, optimizing campaign effectiveness and impact.

Sentiment analysis :
Performing sentiment analysis on reviews(app feedback reviews,call centre feedback reviews) has provided invaluable insights into customer feelings, offering nuanced understanding of their sentiments toward application, products, and te overall user experience.

App Reviews & Portal Reviews Topic Modelling:
Implemented LDA-based topic modeling on portal reviews, uncovering underlying themes and subjects. Identified key topics within reviews, providing comprehensive understanding of user sentiments for targeted improvements.