L K SARMA A

· Spearheaded the design and development of high-throughput Spark jobs for real-time telecom data ingestion and transformation, processing billions of events daily in a microservices-driven architecture on AWS.

· Led the migration of legacy BTEQ workloads from Teradata to modern Spark/Glue-based pipelines on AWS, leveraging S3, Glue Jobs, and Redshift Spectrum for unified data lake and warehousing solutions.

· Designed ontology-driven data models for telecom datasets, enabling quick integration into Foundry analytics pipelines.

· Developed interactive workflow UIs within Foundry Workshop, allowing business teams to manipulate large-scale datasets and visualize trends.

· Built PySpark ETL pipelines to ingest, cleanse, and transform high-volume call data, integrating with Foundry Quiver for analytics.

· Implemented Foundry writeback functionality to synchronize processed datasets with upstream billing systems.

· Created Power BI dashboards for network performance KPIs, leveraging Foundry Contour visualizations for real-time data feeds.

· Collaborated with AI teams to embed natural language processing (LLM) features into customer support tools.

· Used PyTest and CI/CD pipelines for automated validation of ETL pipelines, reducing production incidents by 30%.Ensured end-to-end data integrity and lineage tracking by integrating AWS Glue Data Catalog and schema validation at ingestion.

· Defined scalable DynamoDB designs for storing semi-structured call log metadata, ensuring fast, indexed access for downstream services.

· Designed Java/Python-based APIs for secure, high-concurrency telecom data services, integrated with Foundry writeback pipelines and downstream billing systems.

· Built React/Typescript dashboards integrated with Palantir Quiver/Contour to visualize call volumes, dropped call KPIs, and customer churn insights in near real-time.

· Designed secure data APIs with token-based authentication and integrated them with internal business dashboards to deliver on-demand insights.

· Leveraged Redshift and Snowflake for analytical workloads and built automated data exports into Tableau and Power BI using scheduled Lambda functions.

· Conducted POCs for adopting container-based ML model deployment workflows via SageMaker endpoints and integrated them into production-grade telecom analysis pipelines.

· Engineered scalable, HIPAA-compliant ETL pipelines in PySpark for claims processing, transforming structured/unstructured healthcare data from on-prem to AWS S3 and Glue.

· Integrated TensorFlow-based patient risk scoring models into batch pipelines that auto-triggered based on data readiness, providing actionable insights for case managers.

· Architected near real-time streaming data pipelines using Kafka and Spark Structured Streaming for ingesting high-frequency patient vitals data from external IoT devices.

· Designed secure ontology mappings to integrate patient records across disparate data systems in compliance with HIPAA.

· Built Foundry Quiver dashboards to monitor claims approval SLAs in near real-time.

· Automated ingestion of clinical data from multiple sources, applying cleansing logic in PySpark pipelines.

· Integrated AI-powered analytics for fraud detection, embedding ML models into Foundry workflows.

· Developed writeback capabilities in Foundry to update claim status directly in operational systems.

· Built PyTest suites to validate data accuracy for every ETL stage.

· Developed AIP-powered fraud detection workflows, embedding ML models into PySpark pipelines and surfacing outcomes through Foundry APIs and dashboards.

· Applied Foundry Health Checks to proactively flag anomalies in patient record feeds.

· Worked with frontend teams to improve data visualization usability for healthcare administrators.

· Participated in MLOps initiatives for version control and A/B testing of ML models using MLflow integrated with Foundry and SageMaker pipelines.

· Worked extensively on Palantir Foundry, designing Code Workbooks to transform, filter, and aggregate patient, provider, and claims data; leveraged Ontology to link raw sources with business-contextual layers.

· Used Palantir Contour to develop governed data products with lineage tracking and automated documentation, supporting both audit compliance and reproducibility of analytics pipelines.

· Integrated MLflow-managed models into Foundry pipelines and set up batch inference jobs that triggered based on new data availability, with outputs published directly to Foundry APIs and dashboards.

· Led initiative for implementing schema validation and automated quality checks using Great Expectations framework for claim processing data.

· Integrated Redshift-based analytical reporting with Looker dashboards for executive reporting, with automated refresh pipelines managed through Airflow.

· Collaborated with multiple stakeholders including business analysts, data scientists, and compliance teams to ensure data workflows aligned with federal health data guidelines.

· Took ownership of designing enterprise-level data models, ensuring alignment with business functions such as customer churn, network performance, and call data record (CDR) analytics.

· Architected and implemented robust Azure Synapse pipelines to ingest, process, and transform multi-source telecom datasets including logs, CRM records, and real-time call metadata.

· Orchestrated data movement from Snowflake and on-prem Oracle systems into Azure using hybrid cloud strategies and Azure ExpressRoute for secure, high-speed transfers.

· Designed governance workflows using Azure Purview and Unity Catalog to document metadata, access policies, and ensure regulatory compliance across subscriber data.

· Collaborated with the security and risk team to integrate Azure RBAC and Key Vault, implementing robust data masking policies for sensitive customer attributes.

· Delivered React/Typescript-powered analytics dashboards for end-users, integrating directly with Foundry Contour and Azure Synapse reporting layers.

· Designed CI/CD pipelines with Docker, Kubernetes, and GitLab to automate deployments of ontology changes, Spark pipelines, and dashboard configurations.

· Developed semantic models in Power BI and Tableau, used for self-service reporting on call volumes, latency patterns, and mobile data consumption.

· Conducted intensive training workshops for data stewards and engineers to ramp up on Palantir's Code Workbooks and Object Explorer, helping establish platform-wide best practices.

· Utilized GitLab CI/CD pipelines for versioned deployments of ontology changes, data transformations, and UI configuration artifacts in Foundry.

· Modeled high-volume data for 5G rollout projects by integrating Kafka Streams into Azure Event Hubs, ensuring seamless streaming and batch sync with downstream systems.

· Pioneered the cloud data mesh design by leveraging Azure’s hub-spoke VNet topology, creating domain-specific data products for customer engagement, billing, and service quality.

Developed a custom data lifecycle management solution using Azure Data Lake Storage Gen2 and Azure Functions, automating data tiering and archival processes, resulting in a 35% reduction in long-term storage costs