SPANDAN SANKET MAHARANA

Architected end-to-end ETL pipelines using PySpark and Databricks to process high-velocity insurance records, ensuring real-time data availability for actuarial risk modeling.

Designed a suite of Power BI dashboards with Conversational BI principles, automating performance tracking and reducing manual reporting overhead by 12%.

Developed a broker performance scoring engine by extracting data from Axiom and applying a 4-criteria analytical framework that improved underwriting focus by 10%.

Optimized complex DAX measures and SQL queries, reducing report loading times by 30% for global underwriting teams.

Automated data quality auditing within PySpark ETL pipelines, identifying and flagging 95% of ingestion anomalies before reaching the actuarial modeling layer.

Developed a centralized revenue intelligence system integrating GA4 and MixPanel, providing AI-driven growth forecasting and competitive market insights.

Directed the design of A/B testing frameworks using Experimental Design methodologies to ensure data-driven selection of high-revenue conversion features.

Architected a modular semantic layer for revenue reporting, enabling executive leadership to perform ad-hoc AI discovery and trend analysis without IT intervention.

Directed the migration of legacy user-tracking data to a centralized Cloud Data Warehouse, reducing report generation time from hours to minutes.

Designed and managed a centralized RDBMS (Microsoft Access) to track multi-million dollar research grants and awards, ensuring 100% data integrity for institutional governance and performance tracking.

Developed a Python-based automation engine integrated with Large Language Models (LLMs) to streamline the validation of research datasets, reducing manual audit efforts and strengthening data readiness.

Engineered executive-level Power BI visualizations to monitor departmental KPIs, directly supporting leadership in data-driven budgeting and resource allocation decisions.

Architected a multi-agent Agentic AI decision-support system, utilizing LangGraph and Retrieval-Augmented Generation (RAG) to automate cloud implementation blueprints, reducing manual solution design cycles by 10%.

Engineered an automated Cloud FinOps Intelligence platform with Augmented Analytics, identifying 20% cost variation for executive decision-making through comparative data modeling.

Established an LLMOps monitoring framework using Python to track accuracy and latency of the Agentic AI tool, ensuring model reliability and reducing hallucination rates in cloud blueprint generation by 15%.

Architected a multi-dimensional Star Schema in SQL for 20+ enterprise clients, increasing cross-platform cloud cost data accessibility by 79%.

Implemented automated refresh monitoring and health tracking for cloud expenditure datasets, ensuring 99% data freshness and reducing operational overhead by 15%.

Built a client-facing credit analytics engine by extracting large-scale customer usage data from Azure databases using SQL and Python, generating automated risk exposure reports for B2B enterprises.

Programmed automated reporting workflows using Python and Selenium, reducing manual turnaround time by 20% and increasing responsiveness to Salesforce-tracked client requests.

Collaborated with UX and engineering teams to implement BI-driven UI enhancements for the Credit module, significantly improving data accessibility and system usability for finance end-users.

Optimized project resource allocation by building comparative cost-scope models, improving the alignment between service execution and business outcomes across design and eLearning engagements.