Nov 2024 - Present

Technical Data Analyst at Intuit, Mountain View, CA

• Played a key role in developing and scaling Intuit Assist’s LLM-powered customer support features, enabling chat, voice and video pathways across TurboTax (via Intuit Expert Platform) and QuickBooks (via Salesforce). • Partnered with product managers and data scientists to scope requirements and translate complex technical dependencies into data product needs, shaping timelines, effort estimates and testing strategies. • Engineered data pipelines and clean datamarts (Superglue, SQL, Databricks, Hive) by aggregating clickstream beacon data (Kafka, S3, Salesforce, clickstream datalake, Hive datamarts) across conversation activity, utterances, interactions and transcripts; delivering a unified view at the level of expert, customer, contact, engagement or conversation.

Oct 2020 - May 2024

Senior Data Analyst at Financial Recovery Group (FRG), Fort Lauderdale, FL

•Analyzed a population of 100,000 patients clustered into expenditure buckets. Utilized patient journey (inpatient visits, stays, cost) to formulate recommendations for physicians, focusing on cost control, treatment optimization, and practice pattern adaptation•Performed root cause analysis to identify gaps in data infrastructure, logging & scalability to supplement new data•Analyzed product metrics, including signup rates, customer acquisition costs & conversion rates, using web analytics to drive financial forecasting•Enhanced product engagement through improved UI functionalities, business reviews, targeted in-app messages & feedback surveys, increasing user retention•Automated financial reconciliation process, optimized SQL queries & Python scripts to improve data processing pipelines

Sep 2020 - Oct 2020

Data Analyst at Target, San Francisco, CA

|Demand Forecasting & AI| •Build analytical reports to validate metrics of sales vs forecast values to improve accuracy of demand forecasting engine •Automated existing workflows from schedule based to trigger based events to significantly save time and runtime load •Built greenfield dashboard to analyze 5 department performance forecast values & ad-hoc analysis to resolve inaccuracies

Apr 2020 - Sep 2020

Data Engineer at Kaiser Permanente, San Francisco, CA

•Built 1TB+ Hive big datasets for predictive ML models in Spark to improve efficacy of caregivers to avoid post-visit hospital admissions which led to an increase in bed & resources by 40% for emergency encounters•Created Rshiny & Dash web apps that deliver actionable insights(like cost per treated patient, prioritize disease severity, classify patients into drug class, LOB)to support clinician in reducing unwarranted variation across 729 Medcenter areas•Identified top 5% of social determinant of health risk score of 100,000+ high utilization patients by ad-hoc analysis in HiveQL•Built data pipelines in Azure Data Factory & Synapse to reduced data capture gaps by 30% by mapping ICD 10, CPT, SNOMED•Dropped runtime of data pipeline to <15min that is 65% improvement by enhancements

Dec 2019 - Apr 2020

Data Analyst at Gap Inc, San Francisco, CA

|Real Estate and Store Development Operations| • Built analytical reports by querying Amazon RDS & Amazon Redshift for demand estimation & inventory allocation planning • Identified hotspots for deploying Gap flagship stores by performing geo-spatial analysis to experiment new product launches • Built Power BI dashboard to identify cost recovery opportunities w.r.t design, architecture and construction of flagship store

Jan 2019 - Jun 2019

Data Analyst at Grantham, Mayo, Van Otterloo & Co. LLC, Boston, MA

|Enterprise Data Management|• Reduced data retrieval time to <5min which was a 50% improvement by integrating datasets across multiple platforms • Increased accuracy of pricing funds & assets by 10% by performing ad-hoc analysis to detect fraud transactions • Prevented revenue loss by reducing client dropouts by 20%; built a regression model using time-series & retention analysis • Built stock allocation Tableau dashboard for internal teams; extracted daily data from 40 different data sources; created automated pipeline with different user access levels; reduced headcount for task completion from 40 to 1