Lukas Frei

Lukas Frei is a tech lead in forward deployed engineering at OpenAI, based in New York. His earlier work includes StockStats, an application for building stock-analysis dashboards, and Anpai, a venture he co-founded to analyze meeting patterns from calendar data. His career connects hands-on data analysis, data-tool product work and enterprise deployment.

Learning through working software

Lukas earned a bachelor’s degree in business administration at the University of Mannheim. During a semester abroad in Shanghai, machine-learning lectures introduced him to data science when his programming experience extended little beyond basic SQL. He subsequently attended the NYC Data Science Academy’s bootcamp, learning Python and R alongside statistics and machine learning.

His account of that transition includes a correction to his initial approach: he wished he had spent more time writing programs and less time studying a programming language in the abstract. He recorded mistakes in Jupyter notebooks, building a personal reference from problems he had encountered. He also sought project reviews from instructors and classmates rather than treating a working result as a finished one.

Finance supplied an early application. In 2018, Lukas built StockStats, an R Shiny application for creating customizable dashboards of stock statistics and visualizations. Users could arrange analyses, inspect the underlying data and return to saved dashboards. The project grew from his frustration with the limited statistical tools available on free financial websites. His early project archive also includes collaborative house-price prediction and retail-analysis projects. He collected code for tutorials and analytical projects in his GitHub repositories.

From notebooks to applications and calendar analytics

By 2022, Lukas was a product management associate at Deepnote, writing about the connections between databases, notebooks and data applications. His ClickHouse integration guide traced a workflow from querying a database with SQL to retaining the results as a pandas DataFrame, analyzing them with Python and publishing an interactive application. Parameters let readers explore the results, while scheduled notebook runs refreshed the analysis.

Lukas also co-founded Anpai, a venture focused on meeting productivity. Its open-source anpai R package turned calendar exports into data that could be analyzed with familiar R tools. His introduction to the package explains two complementary parts of that work:

• Making calendars computable: The package reads ICS files into tables, using Rcpp for parsing and fitting the results into tidyverse workflows. Users can calculate meeting statistics and create calendar heatmaps from data otherwise presented as a visual schedule.
• Examining the time between meetings: Alongside meeting totals, the package visualizes breaks in an upcoming week. This makes it possible to examine how the remaining time is divided, as well as how much time meetings consume.

Lukas subsequently joined Palantir as a deployment strategist in New York, following studies at the University of Southern California’s Marshall School of Business. He later moved to OpenAI’s forward deployed engineering organization.