Charitha Sree Veluru

Humanitarians AI · Boston, MA (Volunteer)

• Architected an AI education platform with Python/Flask, LangChain, LangGraph, and LangSmith — multi-agent RAG pipelines over PostgreSQL/PGVector for semantic search across 100K+ research documents at 99.8% content accuracy.
• Built ingestion pipelines transforming 100K+ unstructured research papers into structured course content, reducing course creation time by 95% while serving 25,000+ students.
• Developed a Next.js analytics dashboard with LangSmith agent tracing, achieving a 4.9/5 NPS score.
• Automated AWS ECS deployments via Docker, Terraform, and GitHub Actions CI/CD — 98% uptime, 40% lower infrastructure costs.

ClinicMind · Boston, MA

• Designed ETL pipelines ingesting billing and insurance data from 10+ providers into Amazon Redshift with Apache Airflow and Python, cutting report turnaround from 3 days to same-day.
• Optimized SQL workloads on 500K+ claims records, identifying revenue leakage and reducing claim rejection rates by 22%.
• Engineered a GPU-accelerated claim denial prediction model (XGBoost, RAPIDS cuML, SHAP) with 91% accuracy, reducing denial rates by 18%.
• Deployed a real-time anomaly detection pipeline with TensorFlow Autoencoders, cutting revenue cycle errors by 25% and saving ~$120K annually.

Wipro Limited · India

• Built data pipelines for a medical equipment client, ingesting procurement, inventory, and supply chain data into PostgreSQL — eliminating manual workflows behind 40% of order delays.
• Engineered a multi-warehouse inventory system with Redis and RabbitMQ for real-time sync, driving oversell incidents to 0% and cutting reconciliation from 3 hours to 10 minutes.
• Replaced 4–6s PostgreSQL LIKE queries with Elasticsearch-backed sub-200ms responses across a 10,000+ SKU catalog (30x faster).
• Automated CI/CD with GitHub Actions, Docker, and AWS ECS, cutting 4-hour release cycles to under 18 minutes with 100% success rate.

Bharat Electronics Limited · India

• Analyzed operational and production datasets with Python and SQL, reducing manual reporting effort by 40%.
• Built a full-stack KPI dashboard (React.js, Node.js, PostgreSQL), cutting report turnaround from 2 days to 4 hours.