Skip to main content
← engineering

Jinba Scales Natural Language Workflow Automation for Global Enterprises

By Zero G Talent•

Automating enterprise operations has long required writing custom integration scripts or navigating brittle drag-and-drop interfaces. Jinba is replacing those tools with natural language prompt compilation. The San Francisco startup, backed by Y Combinator's W26 batch, builds an enterprise platform that turns conversational prompts into deterministic, production-grade workflows.

Jinba has quickly gained ground in heavily regulated sectors. The six-person company now serves over 40,000 daily enterprise users across Fortune Global 500 financial institutions, insurers, and manufacturers. To expand the core platform and support enterprise rollouts, Jinba is hiring an onsite team in San Francisco, including a Founding Engineer, a Founding Design Engineer, and a Chief of Staff.

Moving Past Drag-and-Drop Builders

Jinba initially began in 2024 as a traditional visual node editor before shifting course. In an interview on Eight Capital, co-founder and CEO Shoya Matsumori explained that graphical builders introduce usability limits for enterprise staff:

"Traditional no-code low-code tools, you actually need to figure out like the settings by yourself... User actually has like has to have like a mental model of like those settings. In our platform, you don't really have to think about that. Like just ask, 'Hey, like I want to change this to like latest version of the model,' and like AI will just like go through the settings like automatically and it will change the settings."

Instead of direct Python or TypeScript generation, the platform compiles plain-English prompts into structured YAML manifests similar to GitHub Actions. This manifest renders as an interactive graph where workers can tweak parameters through either a visual interface or prompt updates. Restricting workflows to verifiable YAML configurations acts as an internal guardrail, preventing unstructured code execution and maintaining corporate compliance.

The workflows deploy instantly across enterprise infrastructure as REST APIs, Model Context Protocol (MCP) servers, webhooks, or scheduled cron jobs. The platform handles hybrid model orchestration across AWS Bedrock, Azure AI, OpenAI, and self-hosted models, while delivering sub-second workflow execution, multi-tenant isolation, on-premise installation options, and enterprise audit logs.

Founders With Machine Learning and Management Backgrounds

Jinba was co-founded by Shoya Matsumori and Takuya Norisugi.

Matsumori holds a Ph.D. in Computer Science from Keio University, where his research concentrated on machine learning across vision and natural language processing. Prior to launching Jinba, he was a project researcher and research fellow at Keio University, worked as a lead machine learning researcher at PGV, was lead researcher at STANDARD, and founded Carnot and Bluem.

Norisugi serves as Jinba's Chief Operating Officer. He earned a bachelor's degree in economics from Keio University, studied at HEC Paris, and previously worked as a management consultant and engagement manager at McKinsey & Company. His background aligns with Jinba's focus on enterprise procurement, compliance architecture, and corporate sales cycles across banking and insurance.

Automating Regulated Financial Tasks

Jinba's enterprise traction accelerated after a pilot with a major Japanese commercial bank. Matsumori noted that financial institutions frequently spend hundreds of thousands of dollars and months of effort attempting to take internal automation pipelines to production, only to stall on compliance reviews, permissioning, and system integrations.

Common enterprise workflows on Jinba involve repetitive, document-heavy operations such as Know Your Customer (KYC) reviews, commercial loan assessments, and regulatory compliance checks. Because financial institutions already maintain detailed operational manuals and compliance guidelines written in natural language, those rule sets map directly into automated execution graphs.

While traditional software delivery cycles at large banks typically require three to six months to clear audits and deploy to production, Jinba aims to reduce that review and deployment window down to two weeks by using automated risk assessments and fine-grained directory permissions.

Engineering and Operations Roles in San Francisco

With $500,000 in seed backing from Y Combinator, Pioneer Fund, and Palm Drive Capital, Jinba is hiring for three core onsite positions in San Francisco.

On the technical side, the Founding Engineer ($150,000 to $250,000) will build both frontend and backend systems powering Jinba's vibe-coding and workflow automation platform, integrate AI models into natural language-to-action pipelines, and interface directly with enterprise customers.

The Founding Design Engineer ($150,000 to $250,000) is the team's first dedicated design hire. The role owns the complete visual and interactive platform experience, establishing a ground-up design system, developing motion interactions, and writing frontend code using modern React patterns and shadcn/ui.

On the business side, the Chief of Staff ($100,000 to $150,000) will work directly with the CEO on strategic planning, cross-functional operations, and internal execution bottlenecks. The position is structured as a part-time contract role requiring at least 20 hours per week, with an intended transition into full-time employment.

Open Roles at Jinba

Role Location Work Mode Salary
Founding Engineer San Francisco / US Onsite $150,000 – $250,000
Founding Design Engineer San Francisco / US / CA Onsite $150,000 – $250,000
Chief of Staff San Francisco / US Onsite $100,000 – $150,000

See all open roles at Jinba