NILAB NATH

Data Scientist

Ernst and Young GDS

08.2022 - Current

Big Data Engineering & Workflow Automation (PySpark, Pandas, Databricks) Developed scalable data pipelines for validation, preprocessing, and extraction; automated workflows with Pandera schema checks; managed file conversions; implemented data quality modules; and built Plotly dashboards with scheduled pipelines for daily delivery.
Data Warehousing & SQL Analysis (Fact and Dimension Modeling) Performed SQL analysis to identify and map source tables for corresponding FACT and DIMENSION tables.
NLP & Data Privacy Solutions (PII Masking, SpaCy, Transformers, Streamlit) Developed an NLP-based PII masking solution using Transformers and SpaCy, optimized performance with multiprocessing and multithreading, built a Streamlit application for deployment, enhanced compatibility with diverse SpaCy models, and achieved significant reduction in PII detection latency.
Agentic AI & Code Conversion (SAS to Python, Streamlit) Developed an Agentic AI solution for SAS-to-Python code conversion, implemented a compile code node to detect syntax errors; and built a Streamlit-based user interface for the project.
Document Intelligence & AI Process Optimization Developed and enhanced Document Intelligence solutions by generating API payloads aligned with business requirements, fixing bugs, and optimizing workflows. Also developed and deployed Azure Functions for automated solution delivery, and reduced redundant LLM calls to cut processing costs and improve throughput.
SOW Ontology Developed a SOW Ontology engine using SpaCy for POS tagging and LLM workflows to identify key technologies and signatories. Used PGVector for efficient chunk retrieval, automated PDF ingestion with Azure Document Intelligence, and created Delta tables in Unity Catalog. Improved XML creation for PDF forms, streamlining the ingestion process for higher data accuracy and efficiency.

Similar Profiles