Andrew Koulogeorge

I recently graduated with a Master's degree in Computer Science from Carnegie Mellon University. During my time at CMU, I did Machine Learning internships at Pinterest's Advanced Technologies Group (ATG) and AppLovin. I earned my undergrad degree in Math from Dartmouth College (Summa Cum Laude) where I was fortunate to be advised by Soroush Vosoughi.

Experience

• ML Intern @ Pinterest ATG (Fall 2025)
  • I worked on two research questions related to Pinterest's 20 billion parameter foundation model: optimizing the pre-training data mix, and exploring parameter-efficient fine-tuning methods to reduce serving costs without sacrificing performance.
• ML Intern @ AppLovin (Summer 2025)
  • I trained a hybrid Deep Learning Recommendation Model (DLRM) + Implicit Quantile Network architecture for real time impression bid prediction.
  • I deployed an improved model to production during internship; online A/B tests estimated ~$50M/year improvement from impression bid efficiency.
• ML Intern @ harpin AI (Summer 2024)
  • I trained Siamese Neural Nets for customer profile similarity model, removing expert feature creation and enabling customer use cases outside of identity data
• Researcher @ Minds, Machines, and Society (Sept 2023 - June 2024)
  • A. Koulogeorge, S. Xie, S. Hassanpour, S. Vosoughi. Bridging the Faithfulness Gap in Prototypical Models. Insights Workshop; NAACL 2025 (Oral Presentation)
  • W. Ma, H. Scheible, B. Wang, G. Veeramachaneni, P. Chowdhary, A. Sun, A. Koulogeorge, L. Wang, S. Vosoughi. Deciphering Stereotypes in Pre-Trained Language Models. EMNLP 2023
• M&A Intern @ Moelis (Summer 2023)
  • (not a typo) I cooked on PowerPoint, Excel, and financial models for an M&A process (link)