Palash Baranwal

1. Authored a whitepaper establishing Amazon-wide framework for auditing machine learning models 2. Engineered data-driven validation framework for ML model auditing, to evaluate model efficiency and reliability. 3. Developed an LLM-based application to identify Payment Data Elements for Amazon Payments IN 4. Led Amazon-wide PO Splitting efficiency analysis uncovering financial discrepancies via systematic data analysis 5. implemented NLP techniques to successfully identifying unaccounted invoices 6. Conceptualized and executed VizCon, a company-wide data visualization competition during Analyticon 2023 7. Distinguished as top 50 technical interviewer across India for 2023 Student Partner Programs

1. Led the data-analytics on 50+ global audits, driving 200+ findings leading to enhanced security protocols, quantifying significant losses and driving systematic monitoring improvement and strengthened financial controls. 2. Leveraged AWS services, SQL, and Python to detect exposure of critical seller data in Salesforce logs. 3. Analyzed consistent delays in customer-shipments for EDSPs by applying statistical methods to study delivery patterns, uncovering fraudulent collusion between delivery associates and customer accounts. 4. Uncovered process inefficiencies in graded-returns leading to financial anomalies in seller-reimbursements causing losses 5. Transformed manual access reviews into automated solution with anomaly scoring, reducing validation time by 99.7% 6. Analyzed global IT asset management through custom scripting, revealing vulnerabilities in tracking and data-wiping

1. Automated Transshipment Inbound and Route Optimization, Amazon Fresh • Designed an architecture and setup data pipelines to pull data from multiple data sources(MySQL, RDS, S3, User Inputs) • Developed logic to implement features such as Selection Maximization, Inventory Reservation, Demand Redistribution • Collaborated with OR team to build an Operations Research model to generate optimized plan for each zone • Utilized linear integer programming techniques to implement hub prioritization, bin fungibility and intercity transshipments, with monthly impact of xx MM INR. Tools and Languages Used – Python, SQL, FICO Xpress 2. Real time capacity recommendation model, Amazon Fresh • Recommend real time capacity changes based on multiple factors (rate of sale, customer behavior, capacity consumed) • Designed, developed and deployed the model, to maximize capacity to accommodate social distancing constraints • Scaled the model to accommodate 2 hour, 4 hour, staggered and overlapping delivery slots • Tools and Languages Used – Selenium, SQL, Python 3. Store recommendation for a pincode, Amazon Fresh 3P (More, Big Bazaar) • Self-service tool to calculate drive time and distance b/w pin codes and stores for a given traffic model and time of day • Created Optimization model to manage supply and demand, and recommend top 3 stores by distance for each pincode 4. Unhealthy inventory tracking tool, Amazon Fresh • Designed an architecture and setup data pipelines to pull data from multiple data sources(MySQL, RDS, S3, User Inputs • Developed logic to incorporate customer behavior, days of cover, in-stock days, transshipment ARCs • Reduced unhealthy inventory by ~98%. Tools and Languages Used – Python, SQL.