Shashank Kumar

About Shashank Kumar

I am a Data & Analytics professional with 6+ years of experience building scalable decision systems that drive measurable business impact. I specialize in turning ambiguous problem spaces into structured, data-driven solutions using machine learning, experimentation, and advanced analytics.At Amazon (Last Mile Analytics), I work at the intersection of ML, statistics, and business strategy — building forecasting and optimization models, designing A/B experiments that improved product adoption and engineering centralized analytics frameworks that reduced global defect metrics. My work spans demand forecasting (XGBoost), regression modeling, experimentation, BI automation, and GenAI evaluation systems.Beyond tools and models, I see myself as a problem solver who runs toward the fire, not away from it. I thrive in high-ambiguity environments, enjoy owning end-to-end problem statements, and take pride in translating complex analysis into executive-level decisions.My journey has shaped how I think. I began in core R&D at Honda, leading projects for the 2-Wheeler Racing Team. I then spent years preparing for UPSC, building deep cross-domain thinking across economics, policy, ethics, and systems. Today, I bring that structured thinking and resilience into the world of data science — solving real business problems at scale.I combine engineering rigor, statistical depth, and business intuition to build data products that move metrics — not just dashboards.Interested in: Machine Learning, Decision Intelligence Systems, Experimentation Platforms, and AI-driven operational strategy.

Experience

1. Business Intelligence Engineer

   Amazon

   Oct 2023 — Present · Hyderabad, IN

   Engineered a new Analytics Framework by integrating 10+ fragmented data sources to define and develop a centralized Occupancy table calculation for Locker Occupancy and help reduce WW Unavailability defect metrics from 4% to 2% by categorizing into 8 actionable root causes- Led end-to-end design of Partner Engagement Program, developing a scalable weekly performance evaluation framework using statistical thresholding and 12-week lifecycle modelling. Enabled early risk detection and data-driven interventions, delivering $1.2M+ annual savings and improving quality signal accuracy at scale- Developed an automated WW Retail Locker Upsizing Tool using SQL, ETL (Datanet), and QuickSight; identified 2.7K high-potential lockers for upsizing, improved AFN adoption, and saved 400+ manual hours annually through streamlined AWS S3/Redshift data pipelines.

Education

• Jawahar Vidya Mandir, Shyamali

  AISSCE

  2010 — 2012

• National Institute of Technology Jamshedpur

  Bachelor of Technology (B.Tech.)

  2012 — 2016