Manvi Jha profile photo 1 on Tal

Manvi Jha

Data Scientist @ PayPal | Machine Learning • Causal Inference • A/B Testing | GenAI & LLM (Claude / MCP) Analytics Automation | Python · SQL · BigQuery for Payments, Retail & Healthcare | Ex-Uber, ZS

Bengaluru

5+ years of work experience

PayPal
Jul 20251 yr 2 mosPresent

Bengaluru, Karnataka, India

• Built a Claude (LLM) analytics pipeline on the Model Context Protocol (MCP) automating anomaly detection and insight generation across 10 KPIs, 4 channels, and 15+ partners — cut reporting cycle ~99% (3 days → ~5 min) and moved it from biweekly to on-demand for 8+ executives. • Designed a quasi-experimental matched-control study across 12K+ merchants; proved statistically significant TPV uplift (p=0.02) and exposed self-selection bias that reshaped pricing-promo strategy. • Ran YoY cohort/conversion analysis (5 segments, 746K+ merchants, US/global) — found the best migration path (+5.1%) and a high-risk regression (-3.4%) flagged for urgent fix. • Built the OXXO alternative-payment dashboard from scratch, surfacing ~35% unsettled transactions ($1M+/yr) and weak integrations (VTEX 43%, Magento 35% vs Direct 67%). • Re-architected the core PPCP BigQuery pipeline (8 tables, billions of rows), cutting runtime from 4–5 hrs to 30 60 min. • Payroll is Accelon
Logic Intelligence
Nov 20231 yr 2 mosDec 2024

Hillsboro, Oregon, United States

• Built time-series forecasting models + automated pipelines for 30,000+ retail chains, contributing to a 5% revenue uplift. • Applied regression and feature-importance analysis to optimize marketing spend and ROI. • Delivered Power BI dashboards (DAX, row-level security), eliminating 20+ hrs/week of manual reporting; cut data mismatches 70% and led SQL upskilling for 7 analysts.
Uber
Sep 20224 mosDec 2022

Hyderabad, Telangana, India

• Built K-Means and RFM segmentation models on 500K+ users (Scikit-learn), driving +20% Average Order Value. • Designed A/B tests (t-tests, confidence intervals) that drove 15+ merchant adoptions in week one; analyzed clickstream funnels to lift conversion 20%.
ZS
Jan 20193 yrs 3 mosMar 2022

Pune Division, Maharashtra, India

• Developed a Payer–Provider Analog Library in PySpark over 2M+ prescription records to detect prescribing shifts from formulary changes. • Built SQL sales-crediting models tracking 4,500+ reps at a billion-dollar US pharma firm; created COVID-19 impact Tableau dashboards.
National Informatics Centre (MEITY)
May 20182 mosJun 2018

Greater Delhi Area

- Implemented JDBC-based Student Application form using Swing GUI - Designed a sample information portal using HTML
GitHub
@jha-manvi

Manvi’s AI Stack

claude
claude
meta
meta
notion
notion
openai
openai
ChatGPT
ChatGPT
Microsoft Copilot
Microsoft Copilot

Education

Kalinga Institute of Industrial Technology, Bhubaneswar

Bachelor’s Degree, Information Technology

Kalinga Institute of Industrial Technology, Bhubaneswar

20154 yrs2019

Join Manvi and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Manvi Jha on Tal

Headline

Data Scientist @ PayPal | Machine Learning • Causal Inference • A/B Testing | GenAI & LLM (Claude / MCP) Analytics Automation | Python · SQL · BigQuery for Payments, Retail & Healthcare | Ex-Uber, ZS

About

I'm a Data Scientist with 6+ years across the USA and India, turning machine learning, causal inference, and generative AI into decisions that move revenue and product strategy. At PayPal, I lead experimentation and merchant-migration analytics for the SMB portfolio — designing quasi-experimental matched-control studies (12K+ merchants, statistically significant TPV lift at p=0.02), running year-over-year cohort and conversion analyses across 746K+ merchants, and re-architecting billion-row BigQuery pipelines to run 5–10x faster. What I'm most excited about right now is applying generative AI to analytics. I built a Claude (LLM) pipeline on the Model Context Protocol (MCP) that automates anomaly detection and insight generation across 10 KPIs, 4 channels, and 15+ partners — cutting a 3-day reporting cycle to about 5 minutes (~99%) and turning static reports into self-serve, on-demand insights for 8+ executives. Earlier, I built ML customer-segmentation models at Uber (K-Means/RFM, +20% AOV), time-series demand forecasts for 30,000+ retail chains, and large-scale PySpark analytics on 2M+ healthcare records at ZS Associates. Toolbox: Python (Pandas, NumPy, Scikit-learn) · SQL / BigQuery · PySpark · Causal Inference & A/B Testing · Time-Series Forecasting · Claude, MCP & Prompt Engineering · Power BI · Tableau · AWS / GCP I'm open to Data Scientist and Senior Data Scientist roles where experimentation and AI drive the product. Let's connect — manvijha2297@gmail.com

Location

Bengaluru

AI stack

claude

meta

notion

openai

ChatGPT

Microsoft Copilot

Work experience

Data Scientist II — SMB Analytics at PayPal, Jul 2025 - Present · 1 yr 2 mos, Bengaluru, Karnataka, India, • Built a Claude (LLM) analytics pipeline on the Model Context Protocol (MCP) automating anomaly detection and insight generation across 10 KPIs, 4 channels, and 15+ partners — cut reporting cycle ~99% (3 days → ~5 min) and moved it from biweekly to on-demand for 8+ executives. • Designed a quasi-experimental matched-control study across 12K+ merchants; proved statistically significant TPV uplift (p=0.02) and exposed self-selection bias that reshaped pricing-promo strategy. • Ran YoY cohort/conversion analysis (5 segments, 746K+ merchants, US/global) — found the best migration path (+5.1%) and a high-risk regression (-3.4%) flagged for urgent fix. • Built the OXXO alternative-payment dashboard from scratch, surfacing ~35% unsettled transactions ($1M+/yr) and weak integrations (VTEX 43%, Magento 35% vs Direct 67%). • Re-architected the core PPCP BigQuery pipeline (8 tables, billions of rows), cutting runtime from 4–5 hrs to 30 60 min. • Payroll is Accelon

Business Intelligence Lead at Logic Intelligence, Nov 2023 - Dec 2024 · 1 yr 2 mos, Hillsboro, Oregon, United States, • Built time-series forecasting models + automated pipelines for 30,000+ retail chains, contributing to a 5% revenue uplift. • Applied regression and feature-importance analysis to optimize marketing spend and ROI. • Delivered Power BI dashboards (DAX, row-level security), eliminating 20+ hrs/week of manual reporting; cut data mismatches 70% and led SQL upskilling for 7 analysts.

Senior Product Analyst at Uber, Sep 2022 - Dec 2022 · 4 mos, Hyderabad, Telangana, India, • Built K-Means and RFM segmentation models on 500K+ users (Scikit-learn), driving +20% Average Order Value. • Designed A/B tests (t-tests, confidence intervals) that drove 15+ merchant adoptions in week one; analyzed clickstream funnels to lift conversion 20%.

Data Analyst at ZS, Jan 2019 - Mar 2022 · 3 yrs 3 mos, Pune Division, Maharashtra, India, • Developed a Payer–Provider Analog Library in PySpark over 2M+ prescription records to detect prescribing shifts from formulary changes. • Built SQL sales-crediting models tracking 4,500+ reps at a billion-dollar US pharma firm; created COVID-19 impact Tableau dashboards.

Summer Intern at National Informatics Centre (MEITY), May 2018 - Jun 2018 · 2 mos, Greater Delhi Area, - Implemented JDBC-based Student Application form using Swing GUI - Designed a sample information portal using HTML

Education

Bachelor’s Degree, Information Technology, Kalinga Institute of Industrial Technology, Bhubaneswar

Public links

github: https://github.com/jha-manvi