Anthony Berghammer

About Anthony Berghammer

AI/ML engineer building LLM systems, multi-agent architectures, and RAG pipelines…

Experience

1. Research Data Scientist

   RTI International

   Feb 2022 — Present

   Developed and deployed ML ensemble and vector database to classify funding centers from proposal abstracts for large federal client. Built NLP models using Python, Hugging Face, and Ollama, and used LanceDB for semantic search to match projects to funders and maximize grant alignment. Built multiple Monte Carlo-based microsimulations in Python to estimate opioid related outcomes for large healthcare policy clients. Deployed front-end applications using Docker and Streamlit allowing researchers to conduct impact analyses with our models.Developed and deployed hierarchical machine learning text classifiers for federal clients in the criminal justice space using Hugging Face, PyTorch, Docker, and Python, enabling downstream national level reporting on officer-initiated activity. Led team of data scientists to build automated AWS-based data pipelines for wearable sensor data using Python, Docker, and DuckDB. Developed automated compliance monitoring solution to track study participants’ data quality.Built hierarchical unsupervised ML pipeline using Python, Hugging Face, and OpenAI LLMs to generate and analyze a taxonomy of science from dissertation abstracts. Leveraged Sentence Transformers, UMAP, and HDBSCAN for density-based clustering.Led and participated in multiple R&D projects using LLMs including SAS-to-open-source conversion, synthetic data generation, and qualitative coding solutions. Used various technologies, including Ollama, Azure OpenAI, and retrieval-augmented generation.

Education

• Appalachian State University

  Master of Business Administration (MBA), Business Analytics

  2016 — 2017

• Appalachian State University

  Bachelor of Applied Science (B.A.Sc.), Marketing

  2008 — 2012

• Institute for Advanced Analytics

  Master's of Science in Analytics

  2018 — 2019