R+AI 2025 Workshop

From Prompts to Agents: Building & Evaluating R + LLM Workflows

Description

This hands-on workshop is a fast-track for R users who are new to generative AI. We’ll start with reliable prompting patterns in R, then add more context, tool calls, and agentic orchestration to automate multi-step tasks. Along the way we’ll cover tokens & context windows, prompt engineering best practices, model differences, safety basics, testing/QA, and lightweight testing/QC suitable for regulated pharma environments. Demos will feature packages such as ellmer and btw, alongside a small set of generic utilities for HTTP calls, validation, and logging (final package choices may vary). All code and slides will be open-sourced. Participants leave with runnable R scripts that covers a simple agentic workflow plus an evaluation pipeline they can adapt to their teams.

Target Audience & Prerequisites

R analysts/scientists/engineers comfortable with data frames and basic scripting. No prior ML/AI required.

Date & Time

Thursday, November 13, 2025
3:00 PM - 5:00 PM Eastern (2 hours)

Devin Pastoor, PhD

Chief Technology and Product Officer, A2-AI

As CTPO of A2-Ai, Devin works with clients to deliver end-to-end solutions that support quantitative decision-making. Devin specializes in bridging the unique characteristics of each client’s IT structure with the needs of the scientific team. Some of his focuses for clients include providing strategies to migrate from existing tools to more modern cloud-based workflows, and building prototype tools and dashboards to solve business challenges. Before joining A2-Ai, Devin worked as a principal solutions engineer to help dozens of organizations ranging from the FDA to small biotech and device manufacturers to develop and implement

Xu Fei

Senior Solutions Engineer, A2-AI

Xu Fei is a Senior Solutions Engineer at A2-Ai, where he builds AI-powered tools and infrastructure for pharmaceutical research workflows. Working across R and Python stacks, he has developed LLM-enabled applications ranging from interactive chatbots to MCP server implementations, with a focus on making GenAI accessible and practical for scientific computing teams. His work bridges enterprise DevOps, cloud APIs (AWS Bedrock), and domain-specific R packages to help scientists integrate AI capabilities into their existing workflows.

Aathira Anil Kumar

Software Engineer, A2-AI

Aathira Anil Kumar is an AI Engineer at A2-AI and a recent graduate of the University of Georgia, where she earned her Master’s in Artificial Intelligence. With over three years of experience researching and building with AI, she has worked with financial and semiconductor data and now focuses on developing generative AI tools for the pharma world—bridging R and Python ecosystems through open-source innovation and applied research.