P Star Investments
Research across more than 150 technology and AI companies at a family office, and the tools that do the first pass.
Problem
A family office investing in technology and AI has more companies to follow than people to follow them. Earnings, policy changes and funding rounds move prices every day, and most of the first pass (reading, filtering, deciding what matters) was being done by hand.
What I did
I have been an analyst on the technology and AI team since January 2026.
- I research more than 150 listed technology and AI companies, and turn company, industry and macroeconomic news into short recommendations for the portfolio manager.
- I designed and built an automated monitoring system that flags price-sensitive developments. It saves about 10 hours of manual research a week, and it cut the time to a first evaluation by 40%.
- I worked with the investment team to fit the system into how they already work, so the time saved goes to judgement and new ideas.
- I adapted open-source language models to the firm's own research data, which improved the quality of the generated research by 25%.
Evidence
- About 10 hours of manual research saved each week.
- 40% faster first evaluations.
The work is internal to the firm, so there is no public link, and nothing here describes holdings or positions.