Siddhant Madan

• Engineered a Python ETL pipeline integrating Salesforce and QuickBooks APIs to unify data, enabling Power BI tracking of a $50M+ grant portfolio. • Designed DAX-driven Power BI dashboards with drill-through analytics, providing stakeholders real-time visibility into funding trends and proposal submissions. • Optimized AI Document Processing by developing a Python automation script to extract embedded URLs from Word tables and headers, ensuring 100% context retrieval for playbook validation. • Implemented AI auditing workflows using Ironclad (Jurist AI) to automate contract compliance checks, replacing manual reviews and identifying standard-deviation gaps.

1. Interactive Audit Application Client: Cintas Spearheaded a Python/Voila dashboard processing large SKU datasets with 15+ dynamic rules, reducing anomaly detection time by 50% and minimizing supply chain compliance risk. 2. Legal Document Intelligence System Client: HBH Holdings Engineered a resource-efficient RAG system (Haystack/Mistral-7B) achieving 75% accuracy on 250 legal documents, enabling high-performance local deployment and eliminating cloud costs.

Client : Amplify Health (MLOps & Feature Engineering) • Implemented a Universal Data Model (UDM) and tailored Feature Store on Databricks for multi-region claims data, standardizing global analytics and reducing team onboarding time by 40%. • Developed post-processing scripts and feature store functionalities, including data normalization and chronic condition classification, to enhance data relevance for patient risk stratification. • Engineered 50+ reusable ML features using Object-Oriented Python (OOP) and PySpark, delivering specific features for Financial (Regression) and Clinical (Classification) models. • Optimized PySpark transformation pipelines by analyzing execution plans (Spark UI) and implementing broadcast joins/caching, achieving a 30% reduction in system latency. • Formulated a comprehensive data logging framework to ensure traceability and reproducibility for model training, evaluation, and drift debugging. • Resolved complex data quality and model performance issues by collaborating with Principal Data Scientists to build a structured root-cause analysis framework. Client : Jefferson Health (Cloud Migration & Architecture) • Engineered Silver-layer pipelines (Synapse/PySpark) for a strategic migration of 10TB+ Epic Caboodle data to Azure, designing bridge tables and collaborating with clinical coders to validate logic, improving consistency by 25%. • Built a HIPAA-aligned reconciliation framework to enforce schema accuracy, ensuring compliant EHR processing and mitigating regulatory risk across the migration pipeline. • Partnered with the engineering team to refactor legacy SQL logging processes into optimized PySpark DataFrames, resulting in a 50% reduction in notebook execution time. • Supported team delivery by providing technical guidance to junior members on Spark best practices, earning Employee of the Month honors for excellence in team-based execution.

• Analyzed complex CRM datasets to drive customer segmentation, profitability, and personalization strategies, delivering actionable insights through custom web reports and interactive dashboards. • Engineered automated ETL pipelines using Advanced SQL and Python to process millions of records across a multi-tenant architecture, reducing manual reporting overhead by 40%. • Facilitated business requirement discussions and conducted technical feasibility checks to translate stakeholder needs into optimized data models and intuitive user interfaces. • Produced 50+ end-to-end BI solutions for high-scale EdTech and BFSI clients, enabling churn prediction models that improved customer retention by 30%. • Implemented Row-Level Security (RLS) and hybrid data retrieval strategies using Elasticsearch and Redshift to ensure data sovereignty and improve query performance for low-latency reporting. • Collaborated directly with strategic clients to define key metrics for Sales and Marketing funnels, visualizing the end-to-end "Lead Journey" to identify conversion bottlenecks.

Client: The Co-operative Group Project: Retail Data Platform • Orchestrated the strategic migration of a 5TB+ SAP Retail warehouse from Teradata to Azure, leveraging Azure DevOps for CI/CD and Azure Synapse Analytics to ensure data integrity. • Automated the generation of complex SQL scripts using Shell scripting (BASH), which eliminated 2 hours of manual coding effort per day and reduced human error. • Enhanced and optimized existing Linux-based ETL workflows, significantly improving the reliability and throughput of daily retail data ingestion. • Engineered custom Python parsers to digest complex XML, CSV, and SAP flat files, enabling the structured loading of nested data for a high-priority retail business transformation project. • Executed advanced data modeling by developing and maintaining SQL scripts for Type-1 and Type-2 Slowly Changing Dimensions (SCD), ensuring historical accuracy across critical inventory tables. • Optimized warehouse storage efficiency by implementing advanced data compression techniques, reducing the infrastructure footprint while maintaining query performance. • Resolved critical store and customer data discrepancies through deep-dive root cause analysis, utilizing Power BI to visualize data quality trends and report findings. • Analyzed granular basket-level and stock data using SQL to identify supply chain patterns, directly informing optimized inventory replenishment strategies.

Relevant Coursework : Data Visualization | Data Wrangling | Simulation Modeling and Methods | Probability Models | Statistical Methods | Applied Linear Regression | Big Data | Governance of AI/ML Systems | Marketing Foundations