Hi, I am Joachim Talloen

I'm a Data Scientist at McKinsey & Co. leveraging Generative AI and advanced analytics to solve institutional challenges. By utilizing techniques such as machine learning, natural language processing, and econometrics, I advise leaders, expand our knowledge, and develop scalable products and solutions. I provide guidance on pressing global topics, with a particular passion for scaling technological adoption.

Prior to joining McKinsey & Co., I completed a Ph.D. in Behavioral Economics at Carnegie Mellon University, working primarily with my advisor Alex Imas. I also spent considerable time at the Machine Learning Department and the Language Technology Institute, collaborating on research with faculty and taking masters and doctoral-level machine learning and deep learning courses.

Publications

Throughout my time at McKinsey & Co. and Carnegie Mellon I published a myriad of articles in my areas of interest and passion. A selection of the articles I have worked on are listed below:

• To Create Lasting Change, Companies Can Draw on Behavioral Insights
  with Davis Carlin, Kourosh Houshmand, and Sebastian Rodriguez.
  McKinsey & Co., 2023.

• Married to the Job No More: Craving Flexibility, Parents Are Quitting to Get It
  with Aaron De Smet, Bonnie Dowling, and Marino Mugayar-Baldocchi.
  McKinsey & Co., 2021.

• A 680,000-person Megastudy of Nudges to Encourage Vaccination in Pharmacies
  with Katherine L. Milkman, Linnea Gandhi, et al.
  Proceedings of the National Academy of Sciences, 2022, 119(6).

• Megastudies Improve the Impact of Applied Behavioural Science
  with Katherine L. Milkman, Dena Gromet, et al.
  Nature, 2021, 600: 478–483.

Code

I have worked on several interesting AI and data science projects, most notably with Zachary Lipton, Machine Learning professor at CMU and CTO and CSO at Abridge, on the speech-to-text tranformer model Abridge leverages to transcribe and capture noteworthy information from doctor-patient conversations. An interim write up of this work can be found here with the respective python script linked below.

Clinical BERT

Speech-to-text transformer model leveraging clinical BERT with context, in collaboration with Abridge.

Sequence-to-sequence model

End-to-end sequence-to-sequence speech-to-text model with attention.

Multi-layer perceptron

Python code illustrating how to build a simple MLP model using PyTorch.

RCT analysis

Fixed effects and SE clustered regression analyses for an RCT with three conditions.

Fueled by Intelligentsia | December 2025