Shayan <> Twill: technical interview

Date: 2025-03-07 · Duration: 68.56999969482422 min Organizer: Nader Attendees: Martin, shayandashmiz@gmail.com, Nader

Summary

The Technical Interview meeting between Shayan and Martin from Twill began with an introduction where Shayan shared his background in finance, economics, and quantitative engineering, detailing his experiences at Goldman Sachs and a startup, particularly in deploying code and debugging. The meeting progressed to a data analysis exercise where Shayan conducted exploratory data analysis (EDA) on a sample dataset, including examining missing values and plotting correlations, before discussing feature correlation analysis. Martin then introduced language models, focusing on Hugging Face, with a practical component on model outputs and strategies for tokenization. The conversation shifted to the challenges Twill faces in AI engineering, such as generating effective SQL queries, and included an outline of the team structure and the potential for synthetic data generation to enhance model performance. Action items were established for both participants, with Shayan encouraged to continue his AI learning and Martin tasked with considering a RAG pipeline for enriching domain knowledge.

Action Items

• Shayan Continue learning about AI and language models (57:26)

Martin Consider implementing RAG pipeline for accounting domain knowledge (57:26)