Position Title
Agent Harness Engineer
Company
Confidential AI Product Company
Employment Type
Full-time
Location
Nishi-Shinjuku, Tokyo, Japan
Work Style
Hybrid work arrangement with 3 days in office and 2 days remote
Salary
Annual salary: 12,000,000 JPY – 20,000,000 JPY
Monthly salary: 857,143 JPY – 1,428,571 JPY including fixed overtime allowance for 45 hours
Additional overtime paid separately
Stock options available
Performance reviews and bonuses twice per year
About the Company
A rapidly growing AI-native technology company is building enterprise-grade AI systems designed to transform how organizations operate. Founded in 2023 as part of a publicly listed Japanese technology group, the company develops AI infrastructure and products that integrate enterprise SaaS environments and enable AI agents to execute business operations safely and efficiently.
The organization is building the foundational infrastructure layer for production AI agents, with a long-term vision of becoming the operational intelligence layer for enterprises.
Role Overview
The company is seeking an Agent Harness Engineer to design and build the core execution infrastructure powering production AI agents.
This position focuses on the infrastructure layer responsible for orchestration, execution control, memory systems, guardrails, session management, context injection, model routing, recovery mechanisms, and workflow reliability.
This is not a traditional backend API role. The position involves building foundational systems for enterprise AI agent execution at scale, including orchestration engines, SDKs, runtime frameworks, memory architecture, inference optimization, and distributed execution systems.
Key Responsibilities
Design and implement the shared Agent Harness platform across AI products
Build agent execution engines, graph runtimes, and state machine systems
Develop SDKs and internal tooling for AI agent development
Implement session management, checkpointing, recovery mechanisms, and memory systems
Design guardrail and policy execution frameworks for safe AI behavior
Build model routing systems across multiple LLM providers and model types
Develop context management, RAG integration, and memory architecture for production AI systems
Optimize inference pipelines for latency, reliability, caching, and cost efficiency
Build workflow orchestration systems, queuing infrastructure, routing, autoscaling, and batch processing pipelines
Collaborate with research and product engineering teams to productionize new AI capabilities
Partner closely with infrastructure, data, product, and AI quality teams
Support platform reliability, operational monitoring, incident response, and post-mortem analysis
Design data access and permission management systems for secure AI execution
Required Qualifications
Bachelor’s degree or equivalent practical experience in Computer Science, Software Engineering, AI, Machine Learning, Mathematics, Physics, or related disciplines
5+ years of backend engineering experience
Strong production development experience using Python
Experience designing and implementing production systems utilizing LLMs or AI agents
Hands-on experience designing distributed systems beyond operational support responsibilities
Experience designing and implementing RESTful APIs or gRPC services
Fluent Japanese communication skills for technical product discussions
Business-level English communication skills
Preferred Qualifications
Experience with agent frameworks such as LangChain, LangGraph, AutoGen, or similar technologies
Cloud infrastructure and production operations experience on AWS, GCP, or Azure
Understanding of RAG systems, vector databases, and AI memory architectures
Experience with model routing and inference optimization
Foundation software development experience using Go
Strong understanding of Kubernetes and container orchestration
Experience with event-driven architectures using Kafka, RabbitMQ, Pub/Sub, or similar technologies
Experience implementing AI guardrails, policy execution systems, and observability frameworks
ML infrastructure or MLOps experience
Strong English technical communication skills
Technical Environment
Backend / Infrastructure: Python, Go
Frontend: TypeScript, React, Next.js, NX
AI / Agents: LLMs, AI agents, RAG, context injection, guardrails, memory architecture
Infrastructure: GCP, Kubernetes, Docker, Terraform
Messaging: Kafka, Pub/Sub
Monitoring: Prometheus, Grafana, OpenTelemetry
Interfaces: REST APIs, gRPC, SDKs
AI Development Tools: Claude Code, Cursor, ChatGPT, Devin
Team Structure
Engineering organization of approximately 120 members
Cross-functional collaboration with:
Agentic Product Engineers
Research Engineers
AI Quality Scientists
Infrastructure teams
Data teams
Product Managers
Working Environment
Office located in Nishi-Shinjuku, Tokyo
Hybrid work model with flexible working hours
Standard hours: 10:00 – 19:00
Core time negotiable
MacBook with Apple Silicon and dual monitor setup provided
Benefits
Full social insurance coverage
Full commuting expense reimbursement
Stock option program
AI tooling support including enterprise AI development tools
Development allowance up to 30,000 JPY per year
Book purchase support up to 30,000 JPY every six months
Language learning and certification support
Monthly wellness allowance up to 5,000 JPY
Housing allowance up to 30,000 JPY per month for designated areas
Hiring Process
Application review
Coding assessment
4–5 interview rounds
Reference check before final interview
Offer stage