Moshi: An AI tool to turn market data into actionable insights for traders

Anekha Sokhal, a former energy trader at JPMorgan, has launched Moshi, an AI-powered recommendation engine designed to help traders process overwhelming volumes of market data. Speaking at AI Engineer New York, she explains that Moshi merges the functionalities of Bloomberg, AI models like Claude, and recommendation platforms like Netflix to deliver tailored insights.

Moshi processes a wide range of real-time data: futures market time series, technical indicators, news feeds, alternative data sources (social media, podcasts), and private user data (analysts followed, subscriptions). The company initially targets commodity and energy markets, with a processing capacity equivalent to absorbing 300 new pieces of information in four seconds. The founder emphasizes the need to adapt tools to traders' workflows rather than imposing generic solutions.

This initiative addresses the growing challenge traders face in managing information overload, often beyond manual processing capabilities. Moshi aims to automate data extraction and synthesis while remaining flexible to integrate individual user preferences. The project also highlights debates around building models in-house versus purchasing existing ones, as well as the pressure on technical teams to deliver complete systems independently.