• Full-time

Publication date： 2026.09.09

[Fully Remote] AI Product Engineer (ClanExe | AI Partner Platform Development)

  • R&D / Engineer /Technician (AI, Data)
  • Full remote work
  • Salary： 6.5 to 8 million yen
【フルリモート】AIプロダクトエンジニア(ClanExe|AIパートナープラットフォーム開発)
職種Position R&D / Engineer /Technician (AI, Data)
会社概要Company profile The company engages in business activities that bridge the gap between fundamental technology development and practical commercial application, focusing on areas such as industrial applications of image recognition (utilizing deep learning), edge AI development (via network compression technology), and the creation of autonomous agent algorithms for use in drones and other systems. Recognized as a "J-Startup," the company is garnering attention as a Japanese AI venture.
Its product lineup includes solutions for visual inspection via image recognition, autonomous drones, automation (auto-piloting) for construction machinery, HVAC optimization, and the utilization of EEG data; its client base spans diverse industries, including companies such as KDDI, Daikin, Hitachi, Honda, and Toyota.
仕事内容Job description [Product: ClanExe]
Persona Lab’s flagship product is "ClanExe," an AI partner platform.

▼ClanExe Service Introduction Page
https://clanexe.araya.org/

ClanExe is a platform that assigns personalities, roles, skills, and knowledge to AI, enabling users to form teams with them and delegate tasks. Based on the concept of "Forming a party with AI," it allows AI to function not merely as tools awaiting instructions, but as team members with distinct areas of expertise.

Its functionality centers on three key pillars:

Workflow Execution — Designing and executing work processes alongside AI members
Tacit Knowledge Extraction — Converting tacit knowledge buried in the field into explicit knowledge through dialogue
Feedback — Feeding the extracted knowledge back into the field and the organization

Currently, the entry point for enterprise adoption is the "AI Panel," one of ClanExe's core features. It provides a "digital twin" persona pool—built from data on actual consumers and professionals—at a scale of N=1,000. This enables the repeated testing of hypotheses that were previously unfeasible due to time and cost constraints associated with traditional research methods.

Our current business strategy involves establishing touchpoints with client companies through the AI ​​Panel and subsequently expanding to the adoption of the full ClanExe platform. From an engineering perspective, this entails simultaneously cultivating "reliability as a research infrastructure" and "scalability as an agent platform."

[Background of Recruitment]
ClanExe is currently in its early growth phase, having launched only recently. However, Proof of Concept (PoC) projects and business negotiations are simultaneously underway with major corporations across diverse industries—including pharmaceuticals/healthcare, consumer goods, materials/components, and consulting. Consequently, development demands are already extending beyond the product's internal scope.

[Job Description]
As an AI Product Engineer, you will be responsible for the planning, development, and operation of ClanExe features, with a core focus on generative AI technology.

Depending on your experience, your role will center on the following three areas: **Core Product Feature Development**
You will design and implement ClanExe’s core features, including AI agent and persona design (prompt engineering, context engineering), multi-agent dialogue pipelines for extracting tacit knowledge, and RAG/search infrastructure.

**FDE-style Customer Deployment**
For ongoing Proof of Concept (PoC) projects, you will translate client-site requirements into product features and provide customization and implementation support. Working closely with the business development team, you will rapidly incorporate requests gathered during sales discussions into the product.

**Designing Development Processes**
As the product enters its initial growth phase, you will collaborate with the team to define operational monitoring and improvement mechanisms for the SaaS, evaluation frameworks (e.g., LLM-as-a-judge), and workflows that balance FDE-style deployment with core product development.

**■ Specific Responsibilities**
*Note: This list covers a broad range of potential tasks; you are not expected to handle every single one.*

Web application prototyping and production-level development
AI agent and persona design using LLMs (prompt engineering, context engineering)
Design and implementation of multi-agent dialogue pipelines for tacit knowledge extraction
Construction and accuracy optimization of RAG and search infrastructure
FDE-style customization and implementation support for client deployments
SaaS system operation, monitoring, improvement, and inquiry handling
Tech PR activities (blog writing, lightning talks, etc.)
Scope of duties: Development and operation of company services and products, associated tasks, and other duties assigned by the company.

**【Development Environment】**
[Languages/Frameworks] Python, TypeScript, React
[AI/Machine Learning] Anthropic Claude, OpenAI/Azure OpenAI Service, Open-weight LLMs, RAG (Vector DB)
[Infrastructure/Cloud] AWS/GCP/Azure
[Development/Operations] GitHub, IaC, Monitoring tools
**【Team Structure】**
Persona Lab is a small team where researchers, engineers, and business development professionals discuss and collaborate at the same table. Representative Director & CEO Ryota Kanai — Neuroscientist; launched the company after working at the forefront of consciousness research.
Team Leader Taiyo Hamada — Neuroscientist; Principal Investigator (PI) for the Cabinet Office’s Moonshot Goal 9; leads NeuroAI research.
Engineering Team — Responsible for implementing personalities into LLMs and developing infrastructure for digital twin generation; the team to which this position belongs.
Sales & Marketing — Dedicated to lead generation and marketing initiatives.
Business Development — Drives enterprise sales negotiations and Proof of Concept (PoC) projects; serves as the direct point of contact for FDE-style deployments.
応募資格Requirement ▼ Required Skills
3+ years of practical experience in web application development
Experience developing features utilizing LLM APIs (e.g., Claude, GPT)
Experience with development and operations in cloud environments (AWS, GCP, or Azure)
Ability to communicate fluently in Japanese, including technical discussions (target: business-level proficiency or higher)

▼ Preferred Skills
Experience developing multi-agent LLM systems or AI agent frameworks
Experience developing applications utilizing RAG, search engines, or vector databases
Full-stack development experience (primary focus on backend, with some experience in modern frontend frameworks like React)
Practical experience in prompt engineering and evaluation (e.g., LLM-as-a-judge)
Experience building and improving products through direct interaction with customers (e.g., FDE, collaboration with Customer Success)
Experience in product development during the pre-PMF or early stages
Knowledge of human cognition and interaction (e.g., cognitive science, psychology, HCI)
Proactive attitude toward sharing knowledge (e.g., participating in tech communities, speaking at events, writing blog posts)
日本語力Japanese level Business
雇用形態Employment type Full-time
勤務エリアLocation Full remote work
勤務時間Working hours Flextime system (flexible hours: 5:00–22:00)
想定年収Salary 6.5 to 8 million yen
条件・待遇Condition Employment Type: Full-time employee
Probationary Period: 3 months
Estimated Annual Salary: 6.5–8.0 million JPY (Over 8.0 million JPY possible for Tech Lead candidates)
Work Location: Tokyo (Fully remote work available)
Working Hours: Flextime system (Flexible hours: 5:00–22:00)
Days Off: Two days off per week (Saturdays and Sundays), national holidays
Leave: Summer leave, year-end/New Year holidays, and other leave
Overtime: Average of 20 hours per month
Benefits: Comprehensive social insurance, commuting allowance (subject to company policy), and other benefits
選考についてProcess [Selection Process]
Document screening
First interview (coding test)
Second interview (assignment)
Final interview
Comments from United World
back to the list.
TOP