Headline

Currently serving notice period Business Analyst → ML & Product | Turning data into decisions at scale | EdTech & Retail | Python · SQL · XGBoost · GCP

About

I build things with data that actually move the needle. Over 4 years, I've worked at the intersection of analytics, machine learning, and product — shipping models and features that have been adopted by 25K+ students, saved 2,800+ hours of manual work, and driven measurable improvements in retention and revenue. At Allen Digital, I built an at-risk model that flagged struggling students before they dropped off — retaining 5% more students in classes where it was deployed. I also led a database migration that cut infrastructure costs by 80%, and collaborated with product and engineering to ship in-class features now live for 150+ teachers. At Merkle, I scaled analytics coverage from 15 to 1,500 vendor accounts for Home Depot, automated reporting that saved thousands of manual hours, and built a Text-to-SQL bot that reduced data query turnaround from 10 hours to 20 minutes. I'm comfortable in the full stack — from writing SQL and training XGBoost models to defining product requirements and presenting insights to stakeholders. I care about outcomes, not just outputs. Currently open to Senior BA, Product Analyst, and PM roles at companies where data is a first-class citizen.

Location

Bengaluru

Work experience

Business Analyst at ALLEN Digital, Oct 2024 - Present · 1 yr 8 mos, Bengaluru, Karnataka, India, Developed a real-time SQL dashboard to monitor student onboarding, reducing app access turnaround time from up to 15 days to under 24 hours for paid students. Led database migration from Redshift to ClickHouse, cutting infrastructure costs by 80% and significantly improving query performance. Built an XGBoost-based at-risk model scoring 25K+ students on predicted attendance with 89% accuracy — classes using model insights retained 5% more students compared to those without. Defined product requirements for in-class doubts and polls with engineering and product teams, shipping features adopted by 150+ teachers and 25K+ students in live sessions. Analyzed homework availability, submission rates, and feature usage across course tracks — found interactive classes drove 20% lower attendance drop and 2–3% higher renewal rates. Contributed to a team-driven NPS improvement from 40 to 60 and 20% YoY CSAT growth across 300+ centers by identifying gaps in query resolution, classroom experience, and app support. Analyzed mentorship program impact across 20K+ students — mentored students showed 3–5% lower attendance drop compared to non-mentored peers.

Analyst at Merkle, Jun 2022 - Oct 2024 · 2 yrs 5 mos, Bengaluru, Karnataka, India, Trained and deployed ensemble ML models (Random Forest, XGBoost, Gradient Boosted Trees) with SHAP-based explainability to forecast gross sales, average order value (AOV), and marketing channel performance across 1,500+ vendor accounts, achieving 94% accuracy. Scaled analytics coverage from 15 to 1,500 vendor accounts, saving 2,800+ hours of manual analyst work by automating reporting and performance tracking for Home Depot's brand advocate network. Applied ARIMA for monthly demand forecasting, achieving 82% accuracy and surfacing early demand dip signals to enable proactive client intervention. Designed a Text-to-SQL bot using Google's Code Bison LLM, deployed across 1,500 analyst accounts — reducing data query turnaround from 10 hours to 20 minutes by eliminating offshore-onsite timezone dependency.

Education

National Institute of Technology Surat