AI Engineer (LLM Agents & Coding Agents) - Hybrid, Tokyo

cogent labs • Japan

Visa Sponsorship

Apply Now

AI Summary

Join Cogent AI as an AI Engineer to design and build reliable LLM and coding agents, harnesses, and evaluation systems for enterprise Document AI and autonomous software development. Focus on architecture, trustworthy AI outputs, and shipping production-grade products with a founding team. Hybrid role in Tokyo with visa sponsorship and rapid growth opportunities.

Key Highlights

Design coding-agent harnesses, evals, and benchmarks for enterprise AI systems

Work on open problems like autonomous coding reliability and AI evaluation frameworks

Own end-to-end outcomes from design to production with influence over technical direction

Visa sponsorship available for relocation to Tokyo, hybrid work model

Key Responsibilities

Design coding-agent harnesses including context management, tool use, permissions, checkpoints, and tracing

Build specialized evaluation benchmarks with gold cases, adversarial scenarios, and automated validation logic

Develop enterprise AI products combining LLMs with deterministic systems for Document AI and agent workflows

Conduct model experiments (fine-tuning, comparisons, inference economics) when impactful to product quality

Critically evaluate and test new AI research techniques for practical application in production systems

Searching for Development & Programming roles that provide visa sponsorship? Connect with international employers through Development & Programming Jobs with Visa Sponsorship opportunities actively seeking talented professionals.

Technical Skills Required

TypeScript Node.js LLM Agents

Benefits & Perks

Visa sponsorship for relocation to Japan

Paid subscriptions to AI tools (Claude Code, Codex, Cursor)

Compute budget for experiments and fine-tuning

Flexible hours (core: 11:00–16:00 JST)

Hybrid work (Tokyo office + remote)

Explore our comprehensive directory of visa sponsorship jobs from employers worldwide who are ready to sponsor talented international professionals.

Nice to Have

Experience with agent harness design (context, tool use, permissions, checkpoints, tracing)

Benchmark/evaluation set creation (gold cases, grading rubrics, LLM-as-judge, regression testing)

Fine-tuning experience (SFT, LoRA, QLoRA) or small transformer model training

Reproducing/critically evaluating ML research papers

Cloud infrastructure, security, or production monitoring for AI systems

Startup/early-stage team experience

Job Description

Hybrid, Tokyo

Full-time
Founding team
Visa sponsorship available
Early-career to mid-level
Start ASAP

is an AI-native startup spun out of Cogent Labs, building Document AI, AI agents and deterministic systems for enterprises. We're hiring an to work on LLM agents, coding agents and AI evaluation. Coding agents like Claude Code write essentially all of our production code, so our engineers focus on what matters most: choosing the right problems, designing the architecture, building the agent harnesses and evals that make AI output trustworthy, and shipping a product customers rely on.

Want the full job description? Read the complete details on LinkedIn, the original posting.

Continue on LinkedIn

This is a short excerpt. All rights to the full description belong to its original publisher.