Abhishek Garg
Data-driven risk and strategy professional with over 9 years of experience across global payments organizations including Mastercard and PayPal. I specialize in fraud detection, transaction decisioning, and risk optimization strategies. My work focuses on building automated pipelines, reject inference models, and segment-based frameworks to improve approval rates while controlling losses. I collaborate cross-functionally with business, analytics, and engineering teams to deliver scalable solutions and self-service tools. Recently, I have also been working on agentic AI solutions to automate decision workflows, summarize operational insights, and enable intelligent human-in-the-loop responses. I have led end-to-end implementations that reduced manual reviews, enhanced customer enablement, and generated measurable revenue impact. My strengths include analytical problem solving, stakeholder communication, and translating complex data insights into practical business strategies.
REVIEWS FROM CLIENTS
Currently working with Paytm and involved in multiple live projects
Currently working with Paytm and involved in multiple live projects
- Developed RI based methodologies to tag unknown applications to minimize loss.
- Worked on algorithms to optimize existing ...
- Developed RI based methodologies to tag unknown applications to minimize loss.
- Worked on algorithms to optimize existing strategy (ML model cut-offs & rules) to align with monthly business KPIs for Venmo, reducing loss and increasing revenue.
- Developed an automated Revenue forecasting engine using R and Python. Used models like ARIMA, Prophet, TRMF, and DEEPAR.
- ...
- Developed an automated Revenue forecasting engine using R and Python. Used models like ARIMA, Prophet, TRMF, and DEEPAR.
- Developed ML models to generate adversarial attacks and made regression and boosting models robust.
- Created ML models for shifting ATM cash withdrawal to POS transactions.
- Worked on client projects developing ML models for Retail and Restaurants.
- Developed an internal data validation tool, used for anomaly detection.
- Improved Confluence's search engine capability using NLP.