Hire Haritha G. - Vector Database Engineer

Leading end-to-end data and ML engineering for one of the largest pharmacy retail chains in the U.S., supporting clinical decision support and predictive healthcare initiatives across millions of patient records. Built end-to-end clinical/pharmacy ML pipelines using Spark/PySpark + AWS, reducing data processing latency by ~40% and enabling predictive healthcare insights for 8,000+ pharmacy locations Developed NLP extraction pipelines for HL7/FHIR clinical documents, converting unstructured clinical notes into analytics-ready features — cutting manual data prep effort by 60% Implemented LLM + RAG workflows (Transformers + vector DBs: FAISS/Weaviate/Chroma) for semantic retrieval and clinical Q&A, reducing clinician search time by ~35% Deployed scalable inference services using FastAPI + Docker + Kubernetes/SageMaker/Triton, achieving 99.9% uptime for production model serving with full CI/CD Established data quality & compliance validation (Great Expectations/Deequ) across 15+ pipelines, reducing data incidents by 50% and ensuring HIPAA-aligned data governance Integrated model monitoring (MLflow/EvidentlyAI/W&B) to track model drift and performance degradation, enabling proactive retraining cycles Tech: Python, PySpark, Spark MLlib, AWS (S3/Redshift/RDS/Glue/Lambda/SageMaker), Transformers, FAISS/Weaviate/Milvus/Chroma, Airflow/Prefect/Kubeflow, Power BI/Tableau

* Drove data engineering and ML initiatives for a $230B * global retail giant, owning forecasting pipelines and executive-facing analytics across inventory, pricing, and promotions. Built retail forecasting pipelines using Spark/PySpark * Azure Data Factory * Snowflake, improving demand planning accuracy by ~25% across 800 * warehouse locations Developed and deployed ML forecasting models (LightGBM, SARIMA, deep learning) that directly influenced $500M * in pricing and promotional decisions Designed and delivered executive Power BI/Tableau dashboards integrating predictive outputs with retail KPIs, adopted by 200 * business stakeholders Implemented full observability for pipelines and models (Azure Monitor/Prometheus/WhyLabs), reducing pipeline failures by 45% through automated drift-based retraining triggers Partnered cross-functionally with data science, product, and business teams to align ML outputs with category performance goals Tech: Spark, PySpark, Azure Data Factory, Snowflake, Azure Monitor, Prometheus, WhyLabs, LightGBM, Power BI, Tableau

* Contributed to credit risk and financial analytics data infrastructure at a global information and insights company serving 30 * countries and 65,000 * business clients. Built GCP Dataflow * Spark ingestion pipelines that delivered analytics-ready datasets into BigQuery, cutting data availability time from 24 hours to under 3 hours Operationalized real-time anomaly detection and scoring using Python/PySpark with containerized deployments (Docker * Cloud Run/GKE), processing 10M * events daily Implemented data lineage and quality controls using Apache Atlas * monitoring tooling, achieving 98% * data reliability for downstream risk analytics and regulatory reporting Reduced data pipeline incident resolution time by ~30% through proactive monitoring and automated alerting Tech: GCP (Dataflow, BigQuery, Cloud Run, GKE), Spark, PySpark, Python, Docker, Apache Atlas, SQL