About me

I am Samantha Roberts, an AI developer and applied systems builder with a foundation in solid-state physics and nanoscience. I specialize in using Generative AI to solve operational, knowledge, and workflow challenges in complex technical organizations—particularly research facilities and academic environments where expertise is deep but difficult to scale.

My training as a physicist and my role operating a large nanofabrication facility exposed a persistent problem: critical knowledge lives in people, documents, and legacy systems that do not communicate well. Addressing this gap required tools beyond traditional research workflows, leading me to deepen my focus in data science, machine learning, and generative AI. My work centers on building deployable AI systems that improve how organizations capture expertise, support decision-making, and enable people to work more effectively. A core part of this effort is teaching scientists and engineers how to think about AI as an operational tool, not a black box.

For concrete examples of this work, see the portfolio section, which highlights projects spanning Generative AI, automation, and analytics in research settings. Each project illustrates how I approach system design in environments where data, workflows, and institutional constraints are complex.

What I'm doing

  • AI Developer

    Design and build Generative AI applications, agents, and knowledge systems for real-world use. My focus is on practical architectures—RAG pipelines, structured knowledge layers, agent workflows, and user-facing tools—that integrate with messy data, existing infrastructure, and human workflows. I prioritize systems designed for deployment, maintainability, and high-impact use.

  • Facility Director & Professor

    Serve as a professor of nanoscience and director of a shared nanofabrication facility, supporting both academic and industry users. This role grounds my AI work in operational reality—complex equipment, safety and compliance requirements, heterogeneous users, and institutional constraints—and directly informs both system architecture decisions and which problems are worth solving.