TL;DR: Centauri AI is a modern data platform for institutional investors to process and analyze complex investment asset data, starting with Structured Finance. We help investment teams evaluate assets faster and win more deals by turning 2-3 hours of data extracting, mapping, and analysis work under minutes. Contact us here.
Check out our latest demos here:
The Structured Finance industry suffers from antiquated technology, particularly in mortgage securitization. Lenders and investors deal with inefficient processes, using emails and Excel files to exchange complex information on mortgage-backed securities and other assets. This requires manual or outdated software for data mapping and transformation, consuming significant time (2-3 hours daily) even for experienced professionals. Similar inefficiencies are present in other Alternative Investment sectors like Private Credit, Private Equity, and Real Estate, where analysts wade through extensive documents for deal analysis and portfolio updates.
We envision bringing a modern ETL and data science platform built natively with AI to this traditional industry. Traders and analysts can centralize, standardize, and gain insights quickly from all the data they handle without wrangling with copy-pasting between numerous PDFs and Excel files. We believe that equipped with the right tools purpose-built for data in structured finance, AI agents can perform on par or better than humans on these mundane tasks.
Extract
Drop your CSV, PDF, or Excel files into our upload hub. The agent automatically detects the structured tables and indexes unstructured text. Later, these tables are ready for data-cutting and modeling.
Transform
Our AI agent is built for data mapping. It identifies and explains data fields in natural language and writes code to map them to the desirable output format.
Analyze
Our first-of-its-kind analysis agent is equipped with features designed for ad-hoc exploratory analysis. You can auto-run, rewind, or branch out a generative report at any given time to get more insights easily.
Present
We support exporting the transformed data and analysis findings as a nice dashboard or customized in your way, such as in well-formatted Excel sheets. More customized-style exporting is on the roadmap!
He was the co-founder and CTO of Blue Water Fintech, which generated $18M ARR before being acquired in 2022. He experienced the inefficiency of data processing and analytics tasks firsthand when he managed 5 institutional trading books as a portfolio manager.
After obtaining a PhD in Statistics from Stanford, Milan led DS for Pricing and Forecasting, powering over 1 million listings at Airbnb. More recently, at Meta, as a research scientist manager, she led the deployment of Facebook Ads system’s first privacy-preserving AI solution (approx $0.5B+ yearly incremental revenue).
James is a UC Berkeley Alumni in Computer Science and was a Director of Engineering at Tubi, hired the Platform Eng Team managing a >1500c autoscaling K8s cluster and oversaw costs for a $x0M/year tech budget. He was also Principal Engineer at Minted as a full-stack DevSecOps developer.
Together, we’re excited to help bring the modern data stack powered by AI to financial firms.