Devanshi Desai Email & Phone Number

Long Island City, New York, Us

Optimum Media Division Ad Sales and Operations Delivers audience-based, multiscreen advertising solutions to national, local, and political clients.DOMO | Billing Platform | 1 TIM | xg Linear | Strata | Eclipse | Show seeker | Strata

Atlanta, Georgia, Us

❖ As a part of data team, a key contributor in building REPAY’s data warehouse from scratch in AWS to ingest, transform and load 13 distinct acquisition data from their respective legacy platforms to Redshift and S3 buckets.❖ Imposed strict validation rules for data using Python’s data validation libraries such as Pydeeque and Great Expectations.❖ Mapping ETL diagrams in Lucid Charts to plan data migration from legacy systems to AWS, Metabase.❖ Implemented a real-time fraud detection system leveraging Delta tables to ingest, process, and analyze transactional data streams with ACID transaction support, schema evolution capabilities, and time travel functionality for auditing and forensic analysis ❖ Monitored data and ML pipeline’s DAG runs for successful data orchestration in Airflow and Databricks. ❖ Frequently deploying Kafka on Kubernetes for container orchestration and utilizing AWS Kinesis to automate deployment, scaling and operation of container across cluster.❖ Utilized Datahub as a central space for Data Storage, Definition, Delivery, and Governance.❖ Developed REPAY Insight Center (RIC) and REV (REPAY Vault) an internal financial file management systems used by their finance team for their monthly, uploads, edits and downloads.❖ Created User, Groups & Access policies, mapping end-to-end lineage, and empowered data owners to govern their data entities using Tags, Glossary terms, Domains in Data hub and Informatica.❖ Consolidated financial data from different sources for budgeting, forecasting and monitor key financial metrics, track expenses and visualize financial performance trends in Sisense, Tableau and Metabase.❖ Extracted business insights from sales data on Showpad through deep investigation and data modeling.❖ Designed process for paid media data attribution for Marketing Analytics in Hubspot.

San Francisco, California, Us

❖ Part of the Data Analytics and Quality Control team performing data mining on unstructured customer remediation data to find patterns and testing it for ad-hoc and root cause analysis using Python, SQL, Power BI, and SAS❖ Led the identification and execution of customer remediation initiatives across diverse business units, including finance, retail, merchandise, and real estate, ensuring precise data analysis and execution for over 150 cases❖ Developed stored procedures and triggers in SQL to automate data validation processes and ensure data integrity, reducing errors by 60%, also implemented SQL views to create virtual tables that simplified data access and improved reporting capabilities❖ Maintained 20+ Power BI dashboards and reports, providing stakeholders with real-time visibility into the progress and outcomes of customer remediation initiatives❖ Implemented SQL joins and subqueries to combine data from various sources, enabling a comprehensive view of customer profiles and transaction history, and leveraged SQL indexing and optimization techniques to improve query performance and speed up data retrieval for timely customer remediation actions❖ Created secure pipelines for migration of financial data across different locations using Confluence and Data Management system for over 1 million records❖ Conduct regression analysis, decision tree analysis, and other statistical techniques to identify factors contributing to customer complaints and remediation requests❖ Implement machine learning algorithms, such as logistic regression, random forests, and neural networks, to identify high-risk customers and preemptively address their concerns❖ Monitor service level agreements (SLAs), resolution rates, customer feedback scores, and Net Promoter Scores (NPS) to assess the effectiveness of remediation initiatives

Dearborn, Michigan, Us

❖Part of the reporting team delivering user interactive Qlik view dashboards to support Purchasing and Finance Team at FORD Motors to build quarterly accruals for supplier payments.❖ Built data pipelines in Alteryx for a smooth transition of data being acquired from multiple data sources, including SQL server and performed data transformation which includes data wrangling, data migration, data integration, and data warehousing. ❖ Developed 40+ SQL test queries and use cases for 20+ sales performance reports and executed them against SQL server to capture the workflow performance against multiple scenarios by validating the Alteryx workflow and Dashboard against Microsoft Excel reports.❖ Utilizing Scala for developing data processing pipelines, implemented robust error handling mechanism in Scala, enhancing fault tolerance and ensuring the reliability of data pipeline❖ Performed Data Modeling, Query Profiling, and Data Analysis using DAX and SQL❖ Manage end to end complex data migration, conversion, and data modeling and create visualization in Qlikview for finance and purchasing team at Ford.

New York, New York, Us

❖ Developed interactive and automated data visualization reports by collecting data from various sources like Google Analytics, Google AdWords, Facebook, Twitter, etc. and performing data cleansing and using effective data visualization techniques, which gives insights into the performance and allow leadership to make informed decisions.❖ Formulated and documented the physical ETL process design based on business requirements and system specifications with strong ETL design skills, including source to target mappings, transformations, lookups, aggregations, expressions.❖ Architected a data processing pipeline leveraging Google Cloud Composer to seamlessly integrate and automate the lead generation process utilizing Cloud Dataflow for efficient data processing, Cloud Storage for scalable data storage, and Big Query for in-depth analytics, which resulted in a 30% increase in generated leads❖ Developed and monitored over 50 customer lifecycle KPIs (acquisition, engagement, conversion, reactivation, and churn)❖ Conducted meetings with vendors, evaluated and tested their tools, and performed negotiations to reduce cost up to 15%❖ Orchestrated ETL pipelines utilizing Airflow and Python to seamlessly ingest travel destination catalogs, comprehensive itinerary details, consumer data, and daily sales transactions from retail partners into S3 bucket of AWS data warehouse

Hoboken, New Jersey, Us

Capstone Project Client: Greek from Greece (Immigrant-owned restaurant chain) ❖ Part of a noble initiative by Stevens Tech to help Hoboken's local business owners by providing strategies and technical assistance to thrive through post covid repercussions.❖ Partnered with key stakeholders to gather business requirements, created process maps in Microsoft Visio for further implementation.❖ Developed data pipeline to Integrate and automate lead generation process resulting in 30% of increased leads.❖ Led meetings with vendors, evaluated and tested their tools, and performed cost negotiations to reduce costs by up to 15%❖ Designed a custom Tableau dashboard to provide insights and analytical solutions to optimize the marketing campaign❖ Developed and trained machine learning models for classification, regression, and clustering tasks using AWS SageMaker, leveraging algorithms such as XGBoost, Scikit-learn, and TensorFlow❖ Provided data insights and market intelligence through AWS API Gateway and AWS Lambda functions❖ Crafted Unix scripts for file processing – Archival process and Analyzing data❖ Exposure to Oracle database and DB2 databases, several database objects created in databases❖ Created a window spec for the partitions by mentioning partitioning and ordering columns❖ Ranked the validation data frame using Pyspark SQL’s rank function over the window spec, then created new validation dataset by filtering daily orders for each store location, grouping them by store location and aggregating daily order_id❖ Used widgets to dynamically input order_id and filter predicted ratings data frame on that order_id to get recommendations for that store location and taken top 10 orders according to ratings

Surat, Gujarat, In

❖ Transformed 2D websites into 3D websites utilizing A-Frame, JavaScript, and NodeJS.❖ Wrote HTML and CSS scripts to beautify the website and implement watermarks.❖ Created a basic JavaScript form to submit requests for a site visit and cost prediction.❖ Used CSS, JQuery, and JavaScript to beautify the UI.❖ Assisted the marketing team with strategizing and pitching ideas for clients across varied verticals.❖Created tables on MySQL, loaded data from MySQL into HDFS using Sqoop, streaming data to HDFS using Scala and implemented Spark for faster processing, and moved data from HDFS to Hive Integrated Spark with Hive❖ Transformed data using Spark and stored transformed data as parquet files❖ Created tables on Hive and loaded parquet formatted data into them ❖ Performed Hive analytics on transformed data❖ Collected Log data from web servers to integrate into HDFS location.❖ Wrote the MapReduce programs to handle semi-structured and unstructured data like JSON, Argo data files and Sequence files for log data.❖ Migrated HiveQL queries into SparkSQL for improved performance.

Master'S Degree, Information Systems Conc. Data Analytics And Machine Learning

Bachelor'S Degree, Computer Science