Nathaniel Speiser

Worked with cross-functional teams across diverse domains to provide data-driven solutions to clients.

• Contributed to the design and implementation of an LLM-based solution trained on metadata from a data catalog, enabling users to discover relevant datasets using natural language queries.
• Designed and led development of a metadata-driven Databricks ETL pipeline adaptable to multiple client data sources, improving scalability and reducing custom build time.
• Led development of a flexible data profiling pipeline capable of evaluating configurable data quality metrics down to the individual column level
• Built proof-of-concept machine learning and analytics pipelines in Databricks for client demonstrations, accelerating sales and technical onboarding processes.
• Developed and deployed a customizable AWS Glue pipeline using PySpark to transfer legacy healthcare data from source systems to multiple target modules.
• Implemented complex insert/update logic for JSON transaction data ingestion into PostgreSQL and Snowflake, ensuring consistency and auditability across systems.
• Conducted advanced healthcare data analysis using Python, SQL, R, and Tableau to ensure public policy compliance; identified anomalies in risk metrics, uncovering actionable insights for the client.
• Created an internal end-to-end NLP pipeline to analyze financial complaints, utilizing GuidedLDA and CorEx with editable topic seeds for dynamic topic modeling.
• Applied a pre-trained Hugging Face sentiment analysis model to extract nuanced insights from complaint text data.
• Modeled domain entities for a child welfare software platform; built SQL-based reporting tools to enhance team collaboration and transparency.
• Assisted in proposal writing and technical scoping for data platform solutions in support of project bids worth tens of millions of dollars, helping to secure high-value client engagements.

• Selected as a teaching assistant following successful completion of the Metis Data Science Bootcamp, based on performance and communication skills.
• Guided students through the design, execution, and presentation of end-to-end data science projects, providing feedback on modeling, storytelling, and technical implementation.
• Led review sessions covering machine learning algorithms, tools (e.g., scikit-learn, pandas), and industry best practices.

• Analyzed magnetic response data of micron-scale samples using Python to characterize material properties.
• Simulated electron behavior in next-generation angle-resolved photoemission spectroscopy (ARPES) chambers using Igor Pro.
• Conducted ARPES experiments at national laboratory synchrotron facilities to investigate electronic behavior of novel quantum materials.