Data · Analytics · AI Systems
Surabhi Jain
Senior BI Engineer at Amazon building scalable analytics systems, LLM evaluation frameworks, and AI-powered decision tools. 10 years turning data into product impact.
Impact at a glance
Featured Projects
AI Agent and LLM Evaluation Framework
Built a RAG-powered AI agent integrating multiple data sources to self-serve insights via natural language, reducing analysis turnaround from hours to minutes. Developed an LLM evaluation framework that improved response accuracy from 78% to 90% and supported Responsible AI standards.
Marketing Attribution Model
Invented the attribution framework for Amazon's Co-op campaigns — defining Attributed Sales, Lift%, and ROAS from the ground up. Automated reporting cut turnaround from days to minutes, directly improving how Amazon measures and grows brand partnerships
In-Store Ads Datamart
Built a full analytics datamart for Amazon's In-Store Ads from the ground up — dimensional models, fact tables, and Airflow-integrated ETL pipelines. Optimized SQL queries delivered 55% faster runtime and 80% faster ad-hoc processing, directly enabling the Digital Signage Ads reporting launch
Skills & Stack
Experience
- Built data and AI systems spanning analytics, LLMs, and large-scale data engineering, including RAG-based recommendation tools, LLM evaluation frameworks, and optimized data marts using Airflow, SQL, and Spark. Strong in building scalable pipelines and translating complex datasets into actionable business and product insights
- Worked on building scalable data systems and analytics solutions enabling self-serve insights and data-driven decision-making across business and revenue teams. Led initiatives spanning advertiser-to-brand mapping, promotion impact modeling, and data reliability improvements through robust ETL pipelines, CI/CD workflows, and monitoring frameworks
- Worked on marketing analytics and subscription growth initiatives, building attribution models for Co-op campaigns to measure sales lift and ROAS while reducing reporting time from hours to minutes. Also developed an Auto Payments dashboard for Subscribe & Save that surfaced CTR drop insights, led to A/B testing improvements, and contributed to recognition from VP leadership
- Built customer analytics and experimentation systems for Subscribe & Save and Coupons, including a K-means segmentation model that significantly reduced order rejection rates and improved targeting performance. Also drove push notification optimization through A/B testing and cohort analysis, increasing coupon engagement, and automated vendor performance dashboards to streamline reporting and reduce manual effort
- I built scalable data and analytics solutions that modernized legacy systems and enabled faster, data-driven decision-making. My work included migrating large enterprise pipelines to cloud platforms for significant performance gains, developing customer segmentation models for targeted strategies, and designing efficient data models and dashboards that reduced reporting time and improved operational visibility across retail and insurance teams
Get in touch
Open to interesting conversations about data, ML systems, or new opportunities.