Moshi is an AI-native fintech company founded in recent years by Anekha Sokhal, a former JPMorgan energy derivatives trader and Fulbright scholar. It positions itself as a real-time AI recommendation engine (RecSys) for professional traders. Its tech stack integrates futures time series, fundamental data, news, X/Twitter, podcasts, and users’ proprietary data (such as PDF reports), leveraging a self-developed ontology and n-dimensional graph modeling to process roughly 300 new data points every four seconds. The company explicitly rejects the “new Bloomberg” narrative, asserting that it is a hybrid of Bloomberg (market data), Claude (LLM interaction), and Netflix (personalized recommendations), with the key distinction being its focus on workflows rather than dashboards.
Moshi adheres to in-house model development and a closed-loop data system: it employs a self-developed vision-language model to parse users’ uploaded unstructured documents (such as scanned PDFs), while completely avoiding the use of third-party APIs to process sensitive transaction data. On the social data front, it filters out noise by identifying trusted industry sources (e.g., LNG facility operators) and proactively highlights divergence signals—such as discrepancies in vessel or flight count statistics reported by Reuters, Bloomberg, and The Wall Street Journal—transforming these into actionable intelligence rather than pursuing superficial consensus.
The company’s development trajectory reflects a strong customer‑centric approach: all product features originate from in‑depth interviews with real traders, following a “meet first, hire later” philosophy. Currently, the focus is on acquiring clients rather than expanding the team; although recently a single position received 2,500 resumes within four days, Sokhal notes that many of these applications had low signal‑to‑noise ratios, including those from individuals with invalid work visas, covert identity‑verification seekers, and recent graduates lacking practical trading experience. She emphasizes the need to establish a “high‑signal” screening mechanism to distinguish genuine buyers—those with both the intent and capability to purchase—from casual contacts.
Moshi recently drew significant attention at the Finance and AI Cross‑Industry Conference in New York and Houston, with attendees including front‑line traders, AI engineers grappling with the “you are the entire team” mindset, and potential ecosystem partners. Going forward, it plans to broaden its reach through themed webinars and dynamic panel discussions. Notably, the full report does not disclose the product’s launch status, revenue figures, funding amounts, or specific customer names; all technical descriptions and business strategies are based on public interviews and event observations, with key timelines anchored to before October 25, 2026.