Nikhil Kumar

I’m a Data Scientist currently building an agentic Pricing Intelligence platform at WingPoint I2O Retail in Hyderabad. I design systems that turn pricing signals into actionable competitor risk insights—without waiting on manual analysis.

In my work, I architected an RCA engine that classifies price-drop root causes (seller steals, platform defence, retailer matching) with 1P/3P awareness and authorized/unauthorized seller labeling at scale. I also built a production forecasting pipeline on GCP Cloud Run for async execution, cutting manual interventions by 80% and improving forecast accuracy by 30% through ML model ensembles.

I focus heavily on quality and operational reliability. I raised image quality-check accuracy from 60% to 98% using an XGBoost ensemble with optimized feature weighting, fully eliminating manual review. I’ve also automated ad-spend and inventory insight pipelines via RapidAPI and mentored junior engineers on GCP pipeline architecture and production debugging.

Before this, I served as a Subject Matter Expert in DSA at Scaler by InterviewBit, mentoring 250+ students while sustaining a 95%+ satisfaction rate and delivering 2,500+ optimized solutions in C++, Java, and Python with careful edge-case and complexity analysis. Earlier, I built React.js UI components and contributed to CI/CD for automated testing and deployment during a Software Development Internship at Sozo.