Skip to main content
← engineering

OneSchema Builds AI Agents for Enterprise CSV and PDF Workflows

By Zero G Talent•

OneSchema is building software to eliminate manual spreadsheet cleanup and automate file-based data ingestion. Founded in 2021 by Andrew Luo and Christina Gilbert, the 25-person San Francisco company provides automated CSV and PDF workflow tools for enterprise operations, IT, and data engineering teams. Backed by General Catalyst, Sequoia Capital, Y Combinator, Elad Gil, and Contrary Capital, OneSchema raised $6.3 million in seed funding without a pitch deck, as reported by Business Insider.

Today, the company is expanding its core team in San Francisco, hiring for in-person software engineering and go-to-market roles to advance its AI automation capabilities and scale enterprise deployments.

Solving the Unstructured Data Problem

The inspiration for OneSchema came from Andrew Luo's experience as an early software engineer at Affinity, a CRM platform for investment professionals. Speaking in a fireside chat with Product School, Luo explained that onboarding customers required importing opportunity, account, and contact lists exported from existing databases. The engineering team initially assumed an open-source parser could handle file ingestion within weeks. Instead, handling legacy Excel files, hidden formatting characters, and ambiguous date structures required four years of engineering iteration. To manage the ongoing manual work, Affinity eventually hired five dedicated Excel analysts.

That experience led Andrew Luo and Christina Gilbert to launch OneSchema to address the long tail of file-based extract, transform, and load (ETL) workflows. While modern enterprises rely on data warehouses for internal systems, cross-company data transfer still relies heavily on flat files because source and destination schemas rarely match and non-technical teams manage the exchange.

In the Product School interview, Luo described how a major credit bureau uses OneSchema to ingest daily financial and accounts receivable files from thousands of trade group members. Because accounting systems like QuickBooks, Sage, and SAP structure overdue buckets differently, OneSchema normalizes the incoming unstructured files into a single format required to process credit scores.

The company's core platform provides an embeddable CSV importer that replaces custom engineering work. In a video overview published by OneSchema, Christina Gilbert showed how the system uses fuzzy and historical column matching to automatically align file headers with target database schemas, allowing users to correct formatting errors inline. The software also provides downloadable Excel summaries of detected errors and offers more than 50 customization settings alongside inline code validations.

To support automated batch ingestion, OneSchema built FileFeeds. In a product demo from OneSchema, Andrew Luo demonstrated how operations and support teams build recurring ingest pipelines for files received via SFTP, email, or API connections. Users clean sample files in a spreadsheet interface, and OneSchema records and replays those transformations automatically for subsequent file transfers.

Tech Architecture and AI Integration

OneSchema builds its processing services using Rust, paired with React and TypeScript on the front end and Ruby on the backend. The system uses a proprietary in-memory database built in Rust to process files containing millions of rows. Main application data is managed using Postgres, Redis, and Amazon S3.

Rather than placing raw spreadsheets directly into large language models, OneSchema uses artificial intelligence for transformation code generation and recommendation. AI models assist in creating date, address, and text parsing routines that run on top of the company's row-merging and column-splitting engine. The AI system also recommends specific transforms from OneSchema's library based on the data schema detected in incoming files.

Revenue estimates published by GetLatka placed the company's annual revenue at $3 million in 2024 with a team of 20 employees.

Hiring for SF Office Roles

All open positions at OneSchema require five days a week in-person presence at the company's San Francisco headquarters. Compensation includes salary and equity ownership.

The engineering organization is recruiting a Senior Software Engineer ($135,000–$200,000). The engineer will lead zero-to-one development on OneSchema's newest AI agent, partnering directly with Product and Go-To-Market teams to identify enterprise automation opportunities. The role involves designing and implementing scalable infrastructure using React, TypeScript, Rust, and Ruby to process large datasets.

To support customer acquisition, the sales team is hiring a Business Development Representative ($80,000–$100,000). The position is responsible for sourcing outbound sales opportunities, executing metric-driven activity goals, and A/B testing sales messaging to refine the company's playbooks.

Role Location Work Mode Salary
Senior Software Engineer San Francisco Onsite $135,000 – $200,000
Business Development Representative San Francisco, CA, US Onsite $80,000 – $100,000

See all open roles at OneSchema