I'm Simon (he/him), an R developer. I build tools for data scientists at Posit PBC (formerly RStudio).🐛

Most of the time, I'm focused on helping R users get the most out of LLMs:

• tidyverse/vitals: LLM evaluation
• posit-dev/mcptools: Model Context Protocol (MCP) servers and clients in R
• simonpcouch/side: side::kick(), a coding agent for RStudio
• posit-dev/btw: easily provide context on R stuff to LLMs
• simonpcouch/gander: high-performance, low-friction chat for data science
• simonpcouch/chores: an extensible collection of LLM assistants
• simonpcouch/predictive: an agentic frontend for predictive modeling with tidymodels
• simonpcouch/bluffbench: A visual reasoning LLM benchmark
• Positron Assistant: A coding agent for Positron
• simonpcouch/kapa: RAG-based search via the kapa.ai API

Some LLM-related experiments that didn't end up going far:

• simonpcouch/buggy: automatically explain and address errors
• simonpcouch/ensure: automated unit testing for R developers

I previously spent most of my time on R packages for statistical modeling, and still maintain a good few of those packages:

• tidymodels/broom: convert statistical analysis objects to tidy tibbles
• tidymodels/infer: a grammar for tidy statistical inference
• tidymodels/stacks: tidymodels-friendly model stacking and ensembling
• tidymodels/bonsai: model wrappers for tree-based models
• tidymodels/workflows: combine preprocessing, modeling, and postprocessing objects
• tidymodels/tailor: postprocessing with tidymodels
• tidymodels/workflowsets: creating collections of modeling workflows
• rstudio/bundle: a consistent interface for model serialization

Related to the above packages, I also have some pieces of a book called Efficient Machine Learning with R put together.

Another part of my gig is maintaining database interfaces for R:

• r-dbi/odbc: connect to any ODBC-compliant database with DBI
• tidyverse/dbplyr: database backend for dplyr

I also maintain some personal R packages that range in functionality from to performance profiling to data querying to biological methods:

• extendr/mdl: performant reimagining of R model matrices, written in rust
• simonpcouch/syrup: profile memory and CPU usage of parallel R code
• simonpcouch/stopwatch: high precision timings using mocking
• simonpcouch/anyflights: query nycflights13-like data for any recent year and US airport
• simonpcouch/gbfs: query data on public bikes from hundreds of bikeshare programs
• simonpcouch/carpentR: predicting lake algal blooms using plankton dynamics
• rudeboybert/forestecology: methods for model fitting and assessment in forest ecology
• simonpcouch/detectors: prediction data from GPT detectors
• simonpcouch/readmission: hospital readmission data for patients with type 1 diabetes
• simonpcouch/forested: forest attributes in Washington State

Keep up to date with what I'm up to on my website's blog, the tidyverse blog, and the Posit AI Newsletter.