Headline

Analyst @ Flipkart | SQL · Python · A/B Testing · ML · LLMs/MCP | AI Engineering

About

Analyst at Flipkart who turns messy customer and operations data into automated systems, predictive models and decisions that move metrics. What I've built: • Automated RCA framework (MCP + decision-tree logic) for Agent-Serve and Self-Serve: isolates baseline vs mix-shift effects and cut deep-dive turnaround 88% (2 hrs → 12–15 min). • Incident Journey Scoring Model: predicts 0.03%-rare escalations with feature engineering, cohorting and Brier-calibrated scores (75% recall at 10% precision), reducing Escalations Per Million Units by 9pp in two months. • 15+ quasi-experiments and A/B evaluations that improved CTU by 30pp while cutting structural cost; earlier, 3+ A/B tests contributing 100+ Cr incremental revenue. • VOC quantification framework (70% correlation with organic traffic) and a GenAI feedback-classification pipeline with Looker dashboards. • Led the Social Media & Platform Integrity Forum: 150+ non-compliant sellers blacklisted, 7+ executive reviews. Toolkit: SQL · Python · BigQuery · Looker · Scikit-learn · A/B & quasi-experimentation · LLMs/RAG (LangChain, FAISS, HuggingFace) · MCP Open to Analyst, Product Analyst, Data Science and AI Engineering roles. Reach me: kathankathrotiya@gmail.com

Location

Greater Bengaluru Area

AI stack

claude

cursor

gemini

github copilot

hugging face

notebooklm

Work experience

Business Analyst at Flipkart, Jul 2025 - Present · 1 yr 5 mos, Analytics for Customer Support and Self-Serve: automation, predictive modelling and experimentation that cut cost and improve customer experience. • Automated RCA framework: Built an end-to-end Root Cause Analysis framework across Agent-Serve and Self-Serve using MCP, real-time optimised database queries and multi-dimensional decision-tree logic to isolate baseline vs mix-shift effects at granular levels. Cut deep-dive turnaround 88% (2 hrs → 12–15 min). • Incident Journey Scoring Model: Built an ML model to predict rare-event escalations (0.03% base rate) using feature engineering, cohorting and Brier-score calibration. Achieved 75% recall at 10% precision; routing high-risk cases to dedicated queues with proactive outreach reduced Escalations Per Million Units (EPMU) by 9pp over two months. • Experimentation: Evaluated 15+ A/B tests and quasi-experiments where standard split-testing wasn't viable, guiding scale-up/scale-down decisions. Improved the CTU rate by 30pp while cutting structural cost without hurting customer experience. • Platform integrity: Founded and led the Social Media & Platform Integrity Forum with category teams to flag Seller and FSN compliance issues. Created a CEO-tracked charter that led to 7+ executive brand meetings and the blacklisting of 150+ non-compliant sellers.

Event Coordinator at Flipkart, May 2025 - Jun 2025 · 2 mos, ▪︎ Executed end-to-end organisation of ANALYTIQ'25 hackathon for 100+ analysts, managing creative direction, budget oversight, venue coordination, and cross-functional stakeholder collaboration to deliver successful competition. ▪︎ Designed comprehensive event branding and digital presence including 5+ promotional posters, marketing materials, and fully functional event website with traffic monitoring capabilities to optimise participant engagement.

Business Analyst at Flipkart, Jan 2025 - Jun 2025 · 6 mos, ▪︎ Analyzed social media VOC to uncover customer pain points and enhanced case-handling workflows for improved issue resolution. ▪︎ Developed a VOC quantification framework, revealing a strong 70% correlation with platform organic traffic to drive data-backed prioritization. ▪︎ Built a semi-automated GenAI pipeline to classify negative feedback into structured categories, with real-time insights visualized on Looker dashboards. ▪︎ Assessed 3+ A/B experiments, contributing to incremental revenue of 100+ crores by optimizing ad strategy and self-serve experience.

Software Engineer at JubatusAI Labs, May 2024 - Jul 2024 · 3 mos, Ahmedabad, ▪︎ Developed a "Confession" feature within a SaaS application to foster team-building using AWS services, including API Gateway for RESTful APIs, DynamoDB for data storage, and React for frontend development. ▪︎ Utilized AWS Lambda functions to handle backend logic and interact with DynamoDB for CRUD operations on confession data. ▪︎ Deployed a serverless AWS API to support seamless data operations, including transaction history and profile management. ▪︎ Developed a React Native app for loyalty point management in restaurants, including referral rewards via friend referrals and Instagram follows. ▪︎ Designed and implemented OAuth-based Google authentication , integrating Google Analytics into a dashboard for account and property management and scripted API calls featuring OAuth consent handling and JSON file attachments.

Education

Bachelor of Technology - BTech, Computer Software Engineering, Nirma University

12th, JEE Asprint, P. V. Modi School

Public links

github: https://github.com/Kathankathrotiya