MADHUKAR REDDY.
Turning behavioural signals into revenue moves. Engineering the infrastructure that makes analytics actually useful — automated pipelines, live experiments, and systems that don't need babysitting.
analytics → GA4 · Snowflake · Power BI
testing → VWO · 80% win rate
domain → BFSI · CRO · Automation
✓ 3 live automated pipelines
$
Data-Driven.
End-to-End.
Beyond the Dashboard
Started deep in data science — obsessed with what numbers actually say. But an insight sitting in a Slides deck is a dead insight. So I built the pipelines, tools, and systems to make them live.
At iQuanti I own the full loop — funnel analysis, A/B experiment design, automated GA4/VWO → Snowflake → Power BI pipelines, and client-ready tooling that the team actually uses. I don't just analyse; I engineer what comes after the analysis.
Data & CRO
GA4 → Snowflake → Power BI. Fully automated, zero manual exports.
Full-Stack
Production APIs, analytics dashboards, internal tools built from scratch.
Machine Learning
CNNs, classification, and predictive systems built to deploy — not just to demo.
"Insights without implementation are just slides. I build both."
Where I've
Shipped.
Engineering rigour applied to CRO — from experiment design through automated pipelines to custom internal tools that the team actually relies on.
Junior CRO Analyst in the BFSI vertical. I own experiment pipelines, automated reporting infrastructure, and conversion funnel analysis for enterprise banking, insurance, and fintech clients.
Built a fully automated daily reporting pipeline using Supermetrics. GA4 data extracted → transformed → loaded into Snowflake → surfaced in a dynamic client-ready Power BI dashboard. Killed off hours of manual exports per week and modernised the client reporting template from the ground up.
Second automated pipeline pulling VWO experiment data daily into Snowflake via native integration. Rich Power BI dashboard with experiment status views, filtering by variant/date, and comparative performance charts. Zero manual data extraction at any point.
Designed and built an internal web tool for statistical significance calculation and test duration estimation. Ships features the off-the-shelf calculators don't: interactive charts, sensitivity analysis, and stakeholder-ready output. Team uses it daily for experiment planning.
End-to-end ownership of experiments across BFSI clients: ideation → hypothesis → VWO setup → traffic allocation → monitoring → analysis → documented learnings. Consistently high win rates grounded in rigorous pre-experiment funnel deep-dives.
Deep behavioural analysis on enterprise BFSI properties using GA4. Mapped critical drop-off points in loan application, insurance quote, and investment onboarding flows. Turned quantitative signals into actionable UI/UX hypotheses that moved into experiment backlog.
Delivered measurable conversion uplift across banking and insurance properties. Connected CRO hypotheses directly to business outcomes — documented wins tied to revenue impact rather than just significance scores.
Core
Competencies.
Six practice areas. From raw data extraction to live web apps to deployed ML models.
Pipeline construction, behavioural analytics, BI reporting across BFSI domains.
End-to-end automated data pipelines. Zero-manual-touch reporting workflows.
Data-driven web apps, custom analytics tooling, and production-grade APIs.
Full experiment lifecycle ownership — from funnel analysis to concluded learnings.
Predictive models, image classification, and CNN architectures built for deployment.
Dev tooling, version control, and productivity infrastructure.
Systems &
Applications.
Hover a card to flip it and see more. Real projects — data engineering, CRO tooling, web apps, and ML.
GA4 → Snowflake → Power BI
Automated daily monitoring pipeline — zero manual exports, live Power BI dashboard for BFSI clients.
Supermetrics extracts GA4 data daily → Snowflake transform → dynamic Power BI dashboard. Replaced 2+ hours of manual exports per week. Client-facing, production-grade reporting infrastructure.
View on GitHub ↗VWO → Snowflake Dashboard
Daily automated VWO experiment pipeline into Snowflake. Rich filtering and status views. No manual pulls, ever.
Native VWO integration → Snowflake. Experiment results, variant comparisons, win/loss tracking — all live in Power BI. Replaced manual weekly experiment exports.
View on GitHub ↗CRO Calculator
Significance calc + test duration tool built beyond off-the-shelf options.
Interactive significance calculator and test duration estimator used daily by the team. Charts, sensitivity analysis, stakeholder export — features standard tools don't have.
Coming soon ↗IPL Analytics Dashboard
Interactive 2008–2023 IPL dashboard — player stats, team records, match history. MySQL + Plotly + Streamlit.
15 years of IPL data, fully interactive. Player career arcs, head-to-head team comparisons, venue analysis. Built end-to-end from raw CSV ingestion to deployed Streamlit dashboard.
View on GitHub ↗PhonePe PDF Analyzer
PDF statements → structured CSV → interactive spending analysis. Category filters, time-series views, download options.
Parses PhonePe PDF statements into clean structured data. Interactive Streamlit dashboard with spending patterns, category breakdowns, and month-over-month trends.
View on GitHub ↗Dairy Farm Management System
Production-grade FastAPI backend — milk quality tracking, automated billing, multi-role auth. Real API architecture with domain modelling.
Full backend system: fat/CLR quality tracking per batch, automated billing per customer, admin vs farmer role access. JWT auth, RESTful design, PostgreSQL. Built to real production standards.
View on GitHub ↗AI Flower Recognition
CNN classifier hitting ~95% accuracy. TensorFlow pipeline from raw images to real-time inference.
Multi-class CNN — data augmentation, custom architecture, training on imbalanced floral dataset. 95%+ accuracy. End-to-end: raw images → model → real-time inference endpoint.
View on GitHub ↗Coffee Shop BI
Excel BI: pivot tables, KPIs, demand forecasting from transactional data.
Clean Excel BI: KPI dashboards, pivot-based analysis, seasonal demand forecasting from 6 months of transactional data. No over-engineering — just clean analytical storytelling.
View ↗GA4 → Snowflake → Power BI
Automated daily performance monitoring pipeline. Supermetrics extracts GA4 → Snowflake → dynamic Power BI. Killed hours of manual work per week.
End-to-end daily pipeline. Automated extraction, transformation, and loading. Client-ready dashboard with zero manual intervention.
GitHub ↗VWO → Snowflake Pipeline
Daily VWO experiment data into Snowflake via native integration. Rich Power BI dashboard for experiment reporting. No manual pulls at any stage.
Experiment status, variant data, and performance metrics — all live in Power BI, updated daily. Replaced weekly manual exports.
GitHub ↗IPL Analytics Dashboard
15 years of IPL data. Interactive — player stats, team performance, venue analysis. MySQL + Plotly + Streamlit.
End-to-end data project from raw CSV to deployed dashboard. Head-to-heads, career arcs, match history. Full data storytelling.
GitHub ↗PhonePe PDF Analyzer
PDF → structured CSV → interactive spending dashboard. Category filters, time-series, download options.
Parses UPI statements into clean data. Spending patterns, category breakdowns, month-over-month trends in a clean Streamlit UI.
GitHub ↗Coffee Shop Sales Dashboard
Excel BI — pivot tables, KPI tracking, trend analysis, and demand forecasting from raw transactional data. No over-engineering, clean analytical storytelling.
KPI dashboards, seasonal demand forecasting, category performance analysis — all from Excel. Proof that great analytics doesn't require complex tooling.
GitHub ↗CRO Calculator Suite
Custom significance calculator and test duration estimator. Interactive charts, sensitivity analysis, built for analyst-first UX.
Features off-the-shelf tools don't have: sensitivity analysis, multi-variant support, stakeholder-ready charts. Used daily by the team for experiment planning.
Coming soon ↗Dairy Farm Management
FastAPI backend — milk quality tracking, automated billing, multi-role access control. Production-grade API architecture and auth design.
JWT auth, role-based access, fat/CLR quality metrics, automated batch billing. Real domain modelling for a production-intent backend system.
GitHub ↗AI Flower Recognition System
Multi-class CNN in TensorFlow. ~95% accuracy on complex floral datasets. Full pipeline: preprocessing → augmentation → architecture → training → evaluation → real-time inference. Built to deploy, not just to demonstrate.
Custom CNN architecture, training on imbalanced dataset with augmentation strategy. Evaluation: precision/recall per class, confusion matrix. Inference endpoint ready for deployment. The ML equivalent of shipping to production.
View on GitHub ↗GitHub
Activity.
Live from the GitHub API — repos, contributions, and what's currently being built.
Let's Build
Something.
Open to roles in automation engineering, data engineering, or AI/ML. Also up for interesting freelance builds. Drop a line.
When not building pipelines or optimising conversion funnels: