The Recommendation Engine for Commodity Traders — Anekha Sokhal, Moshi
October 22, 2026
AI Summary
5 min readAnekha Sokhal, founder of Moshi, spent her career trading energy derivatives at JPMorgan, where she constantly wished for a tool that could connect the flood of data she was processing into something immediately useful. Now she is building that tool. Moshi is a real-time recommendation engine for commodity and macro traders, designed to surface what matters to a specific portfolio faster than any human could.
Sokhal’s path to founding Moshi was deliberate. After her undergraduate degree in mathematics, statistics, and economics at Warwick, she moved into trading because she wanted a fast-paced, data-heavy career. She saw early that commodity markets were shifting toward a data-driven future. She left JPMorgan with the goal of starting a data company, applied for a Fulbright scholarship—the only one given in the UK for data—and used the interview to connect her trading experience to the energy market crisis unfolding during COVID. That got her to the US, where she worked on machine learning research, published at ICLR with Stability AI, and collaborated with NASA before founding Moshi.
The scale and structure of data
Continue reading the full summary in the app — free to try.
Read Full Summary →Free • No credit card required
Never miss an episode of Latent Space: The AI Engineer Podcast
Get every new episode summarized in your inbox — free, ~5 minutes to read.
No spam. Unsubscribe anytime.
What you'll learn
- 1 (00:00) **Pre-Show & Guest Introduction** - Swati and Anekha discuss the overwhelming inbound from AI Engineers after her talk, and Anekha's background in statistics, math, and trading at JP Morgan.
- 2 (03:45) **From Trader to Founder: The Fulbright Path** - Anekha describes leaving JP Morgan, winning the Fulbright scholarship to come to the US, and the genesis of Moshi.
- 3 (05:46) **What is Moshi? The Product Vision** - Anekha defines Moshi as the program she wished she had as a trader, providing real-time insights for commodity and energy markets.
- 4 (08:57) **The "Netflix + Bloomberg + Claude" Analogy** - Anekha rejects the "New Bloomberg" label, instead positioning Moshi as a recommendation engine for traders.
- 5 (11:10) **The UI/UX Challenge: Push vs. Pull** - The conversation shifts to the critical challenge of getting users to actually use the software, comparing it to Spotify and Instagram.
- 6 (13:19) **Alternative Data & Trust: Lessons from the Buy Side** - Swati shares his experience buying expensive alternative data (like the Twitter fire hose) and questions how to filter signal from noise and ensure data privacy.
- 7 (16:38) **Build vs. Buy: The Vision Language Model Decision** - Anekha explains why Moshi builds its own vision language model for processing user data rather than using a third-party vendor.
+ Full timestamped outline available in the app
Show Notes
Anekha Sokhal, CEO of Moshi, joins swyx to discuss building a recommendation engine for commodity traders, turning market data into useful signals, and designing AI that fits users’ workflows.
Watch: YouTube · Moshi · Anekha
0:00 AI Engineer New York and “you are the team”
0:54 Meet Anekha
2:09 From mathematics to energy derivatives
3:46 Fulbright and becoming a founder
5:45 What Moshi does
6:10 Market data
9:03 Bloomberg and market workflows
10:03 Bloomberg, Claude and Netflix
12:03 From dashboards to workflows
15:03 Trusted sources and signal
17:13 Models and document extraction
18:20 AI Engineer New York
20:32 Customer meetings
23:54 Hiring and customers
25:55 swyx’s path into AI
This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit