Posit

Regeneron's Journey to 20X Mentor-Mentee Matching with Posit Team

Regis A. James, Associate Director of Data Science at Regeneron shared his experience working with Posit. Their team uses open source data science and Posit to build their MAGNETRON AI matchmaker platform that automatically connects mentors and mentees to share knowledge across the Fortune 500 company. Pictured below is the team bringing this idea to life: Regis A. James, Jordan Leao, and Jennifer (Galarza) Michel.

Expanding access to growth opportunities across the company:

To facilitate mentorship, the Regeneron team had previously paired people manually with other employees to help them learn about the business and advance their own careers. The process was time consuming and fraught with data quality issues. One cohort received 140 applications, and it took administrators 3 weeks to process them and assign matches. At this rate, extending the program across the company was not possible

The people-ops team turned to the work that was being done in other areas of the business with data science to accelerate data-driven decision making at scale.

Utilizing Posit Team, Regeneron successfully developed an in-house AI platform tailored for personalized matchmaking, effectively removing past obstacles and facilitating its expansion to benefit a broader group of employees. The first cohort of this new tool received 400 applications that were successfully matched within 3 days. This was a 20X acceleration from the previous manual process. Since then, additional matching cohorts have been opened up to over 4000 people across the company.

Quick look into the tools bringing this to life:

To transform months of manual collaborative work into days, Regis A. James built the MAGNETRON AI matching platform using a combination of Posit, R, Python, Large Language Models, Shiny, MySQL, Neo4J knowledge graphs, JavaScript, CSS, HTML, and Bash.

Specific packages with a special mention include: reticulate, bs4Dash, dt, plumber, dbplyr, and neo4R.