Theja Vadla

Data Engineer @BlackRock

NY, US

Designed and deployed 15+ scalable data pipelines on Azure Data Factory, Databricks (Delta tables), and Snowflake, processing petabytes of data and reducing ETL runtime by 40% while optimizing cloud costs.• Developed deep learning models (LSTM, Transformer architectures) in Python (TensorFlow, PyTorch) for financial time-series forecasting, improving risk prediction accuracy by 28%.• Optimized Snowflake-Spark integration using Snowpark API and query pushdown techniques, accelerating complex portfolio analytics by 60% through intelligent caching and optimized DataFrame operations for risk calculation workloads.• Built LLM-powered data extraction solutions using LangChain, OpenAI APIs, and RAG pipelines, automating analysis of 10-Q/K reports and earnings transcripts with 92% accuracy, reducing manual review time by 70%.• Automated ML model CI/CD workflows using MLflow, Kubeflow, and containerized deployments, reducing failed releases by 75% and enabling hourly model updates.• Standardized data and model documentation workflows, collaborating with MLE and Security teams to ensure 100% regulatory compliance and reduce approval time by 50%