Harish

Identified data quality issues within the Digital Platforms.
Completed data lineages.
Implemented data controls to enhance digital growth and innovation.
Identified critical data elements.
Executed data quality assessments, resulting in 20% improvement in data governance.
Streamlined data management processes resulting in 20% increase in data accuracy and 15% improvement in pipeline reliability.

- Identified data quality issues within the Digital Platforms, completed data lineages, and implemented data controls to enhance digital growth and innovation.
- Identified critical data elements and executed data quality assessments, resulting in 20% improvement in data governance.
- Streamlined data management processes resulting in 20% increase in data accuracy and 15% improvement in pipeline reliability.

Worked as Data Engineer across four digital products: Payroller, Bookipi, Bookipay and eSign.
Advised as the Subject-Matter Expert in the Data Team.
Developed processes and adopted Industry Standards/ Frameworks to handle millions of rows at terabyte scale for operational application and various events-based systems.
Proposed a long term Strategic Big Data road-map for Bookipi group and successfully convinced the senior management to invest in Big Data Warehouse-Big Query.
Integrated Production Database (MongoDB) with BigQuery via Hevo Data without any transformations in the ETL process by dumping raw data into a staging area (dataset/schema).
Successfully introduced data manipulation via dbt (data build tool) with Big Query and DataProc (SQL and python/ pyspark) clients, and fully integrated dbt-core with CI/CD setup via Cloud Build, Cloud Source Repository, Artifact Registry and Cloud Run.
Wrangled and analysed data at terabyte scale via optimising the data infrastructure.
Broke raw tables down to construct Multi-dimensional OLAP models.
Enabled appropriate testing on critical transformational stages via dbt-test for pipeline reliability and Data Controls.
Improved the pipeline performance and reliability by 30%.
Created a series of abstractions (views and macros) to reduce the complexity of models and finally materialized tables for reporting tools consumption.
Improved the query performances by 90%.
Delivered simplified metrics tracking report via Google sheet for financial modeling.
Set up end-to-end industry best practices for building an Advanced Data Architecture.
Resulted in 100% automation and self-service.
Integrated GA4 data to BigQuery to understand signup source.
Created a Markov Chain Model by mining and linking it with the operational database.
Resulted in ~65% of matching user_id and user_pseudo_id.
Successfully introduced DataHub for data dictionary, metadata, data linage and management control.
Established an End-to-End data systems for Reverse ETL via Census data.
Worked on metrics alignment across the teams and campaign automation.

Ran a quarterly campaign to move existing customers to eStatements from paper mail.
Provided customer analytics (segmentation and forecasting) pre-deployment to tailor marketing creatives.
Performed post-implementation analysis to understand the campaign’s efficacy.
Targeted 101k accounts with a migration rate of ~80% across the segments for savings and transactions accounts.
Resulted in cost saving of $228k.
Created a prototype for automating the Optout Business Process via Python.
Reduced DigiMails campaign build time by 20%.

Worked primarily on EY's Capital Edge platform (fully integrated Azure blob storage, databricks, ADX, ADF, SQL Server, PowerBI).
Created a forecasting model for a property management client.
Built a model that compares actual, budget and forecasted Cash Flow for month-on-month property Funding, Construction, Rental, Operational, Tax, Asset and Financing cost.
Created an end-to-end solution design of a forecast model.
Completed and delivered construction module using Pandas, Numpy and Koalas via package wheel to work with databricks.

Worked across internal departments, including Enterprise Information Systems (EIS), Project Management Office (PMO), and Global Supply-Chain (GSC).
Increased Callaghan Innovation funding grant by $1.5 Million and a total claim of $3.75 Million.
Designed and established Supply-Chain, Production, Rocket Design Change Management KPI Reports and Dashboards for senior management (Directors and Vice Presidents).
Worked with stakeholders throughout the organization to identify opportunities for leveraging company data to drive business solutions.
Worked closely with Data Operations team to establish and standardize ETL pipelines for Advanced Analytics Ecosystem.
Mined and analyzed data from structured and unstructured databases to improve process, formulate techniques and establish best practice methodologies to ensure operational success (RDBMS, NoSQL).
Assessed the effectiveness of new data sources and conduct feasibility studies and ETL pipelines.
Assessed, deployed, and standardized Advanced Analytics reporting tools and mentored graduate, and intern analysts.
Understood business problems and found appropriate custom ML algorithms for a robust and scalable solution.
Used Advanced Analytics techniques to identify patterns and trends to optimize internal operational processes, revenue generation, waste reduction, planning, and other business outcomes.
Developed processes and tools to monitor and analyze the data accuracy, model’s performance, data drift.
Successfully created a Data Lake for reporting purposes and assisted in Power BI deployment across the business.
Created an End-to-End Advanced Analytical reporting solution (Jupiter notebook) with configuration management (gitlab).
Designed rocket costing model alongside the finance team.
Identified and resolved production bottlenecks by providing innovative analytical insights.
Discovered the root cause of parts shortages through analysis of Bill of Material (BOM) changes.
Identified reporting error due to polluted datasets and developed anomaly detection algorithms to rectify and clean data, by monitoring.

Responsible for the development of a risk prediction model (estimating loan delinquency migration and the probability of defaults) by using continuous time-varying features.
Identified and formulated ~600 time-varying attributes by mining datasets and examining customer spending behavior for portfolio risk management.
Reduced high-level time-varying attributes by Principal Component Analysis, Lasso, and Recursive Feature Elimination.
Validated the reduction with Pearson Correlation, chi-Squared and other statistical techniques.
Data mined validation set from historical customers addressed the problem of bias classes response by experimenting with Under-Sampling, Over-Sampling, and Ensemble classifiers using samplers internally.
Experimented with various pre-processing and Machine learning models, such as eXtreme/ Light Gradient Boosted tree classifier, Gradient Boosted tree classifier, Random Forest classifier (Ensemble algorithm), Elastic-net classifier, generalised, generalised additive, Regularised Logistic Regression (L1 and L2), Support-vector Machine (SVM).
The recommended model outperformed the current risk prediction by ~15% Accuracy and 10% Sensitivity.

Performed data management, statistical analysis, and process auditing for Environmental, Quality, Health & Safety Management Systems (ISO 14001, 9001, 45001).
Coordinated across internal teams to ensure all necessary data were captured and monitored according to the Standard Operating Procedures.
Produced training manuals and led staff training.
Conducted various process improvement audits.
Optimized ARR (Action, Review & Recommendation) process and implemented corrective actions that resulted in external audits with no prominent or minor non-conformances, based on local Regulatory Requirements.
Used statistical analysis to reduce the sample size for several QA processes.