Trilochani Kadali

As a Senior Data Engineer at Citi Bank, I led the design and development of a Customer Spending Analytics Platform that centralized multi-channel transaction data into Azure Synapse Analytics, Azure SQL Database, and Azure Data Lake Storage. I built serverless ELT pipelines using Azure Data Factory, Databricks, PySpark, and Azure Functions, automated ingestion from diverse sources, and developed reusable Spark frameworks for high-volume transaction processing. I engineered data transformation and enrichment workflows, ensuring data quality with Python, Azure Data Quality Services, and Informatica IDQ, while enforcing compliance through Azure Active Directory (AAD), Key Vault, and audit logging. My work enabled personalized customer dashboards, branch-level analytics, and marketing insights, reducing pipeline latency and driving data-driven engagement and revenue growth. Tech Stack: Azure Data Factory, Azure Synapse Analytics, Azure Databricks, Azure Data Lake Storage, Power BI, Azure Functions, SQL, Python,Db2, datastagem, ML, PySpark, Informatica IDQ, Azure Purview, Key Vault, CI/CD (Azure DevOps), Unix/Linux Shell Scripting, Data Governance, Data Lineage, Data Quality.

I led the design and implementation of a centralized Clinical Trial Data Platform on AWS. The project addressed fragmented data sources by utilizing Amazon Kinesis and AWS Glue to process over 50 terabytes of trial data, reducing the time from collection to analysis from weeks to hours. I architected a multi-layered data lake on Amazon S3 and loaded clean data into Amazon Redshift, which accelerated analytical queries by 60%. This work reduced data-related errors by 40%, ensuring 100% HIPAA compliance and enabling faster, more informed decisions for research and drug development. Tech stack: Data Ingestion & Processing (AWS Glue, PySpark, Python, Amazon Kinesis), Data Storage Warehousing (Amazon S3, Amazon Redshift), Data Governance & Security (AWS Lake Formation, AWS IAM, HIPAA), Business Intelligence (Amazon QuickSight), Automation (AWS Lambda, ETL).

The project centered on building a customer behavior analytics platform using a hybrid multi-cloud strategy (GCP and AWS). The primary challenge was integrating fragmented customer data from website clicks to in-store purchases into a unified view. This was solved by creating a streaming data pipeline with Google Cloud Pub/Sub and Dataflow and building ETL jobs on AWS Glue and Lambda to process data efficiently. The outcome was a centralized data model in Google BigQuery and Amazon Redshift that enabled personalized recommendations and accelerated marketing insights, ultimately driving a 15% increase in online conversions and reducing data latency from hours to minutes. Tech stack: Data Ingestion & Processing(Apache Kafka, Apache Spark, Python, Apache Druid, Apache NiFi), Cloud & Storage(Google Cloud Storage, Amazon S3, Google BigQuery, Amazon Redshift, Teradata), Workflow & Automation(Apache Airflow, Jenkins), Business Intelligence & Analytics(Tableau, Power BI, Impala, dbt), DevOps & MLOps(Docker, Kubernetes, GitLab CI/CD).

Led the modernization of First Citizens Bank’s data infrastructure on Azure, building a fraud detection and customer analytics platform. Engineered streaming pipelines with Azure Event Hubs and Databricks to process transactional and behavioral data, enabling proactive risk management and personalized insights. Designed and implemented a full MLOps pipeline with AKS and MLflow for automated model deployment, monitoring, and retraining. Built a centralized data lake and warehouse on Azure Data Lake Storage and Synapse Analytics to unify multi-source data for analytics and reporting. Established robust data governance, monitoring, and CI/CD processes, ensuring compliance, operational efficiency, and high-quality insights for business stakeholders. Teh stack: Data Processing (Apache Kafka, Spark Streaming, Apache Spark, Python), MLOps & Deployment (Kubernetes, MLflow, Docker, GitLab CI/CD), Cloud & Storage (AWS S3, GCP Cloud Storage, Amazon Redshift), Analytics (Scikit-learn, Prometheus, Grafana), Orchestration & Automation (Apache Airflow).

I was responsible for designing and implementing ETL pipelines to ingest, transform, and load petabytes of structured and unstructured data. This involved using a Hadoop ecosystem to handle the distributed storage and processing of massive datasets. I developed custom data ingestion scripts and Spark applications to cleanse and aggregate data, ensuring a single source of truth for the business. The platform's success was measured by its ability to reduce data processing latency, provide insights, and enable the marketing team to launch more effective, targeted campaigns.