Forward Deployed Engineer

Kurashi Tech and the FDE role

Kurashi Tech rebuilds business operations with AI and delivers value through three pieces.

• Caiva: an AI-readable knowledge base that combines industry and domain expertise with customer documents to build a knowledge space AI agents can use
• SHISHI: an engine for embedding AI into operations, with vibe-coding features that generate and adjust workflows and skills from natural language instructions
• FDE: a person stationed at the customer site who uses Caiva and SHISHI to turn real operations into AI-driven workflows

How this role differs from traditional engineering

This is not a job where you simply write backend code in Python.

Because SHISHI's vibe-coding features and Caiva's knowledge base minimize coding from scratch, FDEs need these abilities more than raw coding volume.

• Structure operations: observe and interview customers, then break work into forms AI can handle
• Give good instructions to AI: guide SHISHI precisely and extract the workflows you need
• Design knowledge: decide what knowledge goes into Caiva, at what permission level, and in what structure
• Persist on site: operations do not follow textbooks, so you must work through exceptions until the system really works

Responsibilities

• Observe and interview customer processes and identify what can be AI-enabled
• Implement consultant-designed workflows on SHISHI
• Load customer-specific knowledge into Caiva and design permissions, classification, and update flows
• Build and tune workflow skills and agents through SHISHI vibe coding
• Connect ERP, chat, email, and other business tools
• Evaluate accuracy, tune behavior, and handle exception cases
• Extract reusable skills and patterns across customer projects
• Feed feature requests back to the SHISHI and Caiva teams

Requirements

Candidates should meet all of A, B, and C below.

• A. Can structure operations: observe complex business processes, organize them, and draw workflow, data flow, and permission designs
• B. Uses AI in real work: uses ChatGPT, Claude, Gemini, or similar tools daily and can give precise instructions; can adapt to Claude Code, Cursor, and other AI development tools
• C. Can work with customers: native-level Japanese, ability to go on site and speak directly with operators, and honesty to say what you do not know
• Plus one of the following: web app development experience, SIer or IT consulting experience, RPA/iPaaS/no-code experience, or data analysis experience

Preferred

• Experience building LLM-powered business apps or agents
• Enterprise project experience
• Domain knowledge in finance, healthcare, manufacturing, professional services, real estate, or similar fields
• Basic understanding of TypeScript, Python, or SQL
• English communication ability
• Customer negotiation, presales, or consulting experience

Who this fits

• People who seriously want to change work with AI
• People more excited by untangling and rebuilding operations than by coding itself
• People willing to persist through gritty work at customer sites
• People who enjoy using AI tools heavily and trying new methods

Selection process

1. 1Casual interview
2. 2Document screening
3. 3First interview with current FDE members
4. 4Assignment session: interview a mock customer and design how to implement with Caiva + SHISHI
5. 5Second interview with consulting and development leaders
6. 6Reference check
7. 7Final interview and offer