S

Sakana AI

Orchestrates multiple LLMs via API

Software Engineer - Platform

Platform EngineeringConfirmed live in the last 24 hours
Full-timeMinato City, Tokyo, JapanNo salary listed

About the Job

Requirements

  • At least 3 years of professional software engineering experience in platform, infrastructure, or site reliability engineering.
  • Backend development experience with at least one of Python, TypeScript, or Java.
  • Experience designing and operating production systems using container orchestration such as Kubernetes.
  • Experience building systems on public cloud platforms such as AWS, Google Cloud Platform, or Microsoft Azure.
  • Practical development and operations experience using continuous integration and continuous delivery and infrastructure as code such as Terraform.
  • Experience developing applications or infrastructure using large language model application programming interfaces such as OpenAI, Anthropic, or Google.
  • A product mindset focused on designing and developing from the user experience.
  • Strong motivation to improve development productivity and automate processes.
  • Ability to communicate effectively in business settings in both English and Japanese, including business conversation and document writing, and to adapt language use to the situation.
  • Japanese proficiency equivalent to Japanese-Language Proficiency Test N2 or higher, or equivalent practical experience.
  • English proficiency equivalent to Eiken Grade 2 or higher, or equivalent practical experience.

Responsibilities

  • Design, develop, and operate Sakana Aquarium.
  • Collaborate with Solution and Product teams to integrate Platform components into solutions and improve development efficiency and quality.
  • Build and operate infrastructure supporting production, including continuous integration and continuous delivery, monitoring, scaling, and security.
  • Provide software development kits, templates, documentation, and agent skills for Platform developers, and support developers.
  • Collect feedback through collaboration and proactive user interviews, and incorporate it into platform improvements.
  • Collaborate within a global team using both Japanese and English.

Desired Qualifications

  • Frontend development experience using React, Next.js, or TypeScript.
  • Experience with artificial intelligence agent frameworks such as LangGraph, Mastra, or OpenAI Agents SDK, or with Model Context Protocol.
  • Experience designing and building observability infrastructure using Datadog, Prometheus, or OpenTelemetry.
  • Experience operating large language model operations, including prompt management, evaluation pipelines, or tracing infrastructure such as Langfuse or MLflow.
  • Experience designing large language model evaluations, including evaluation datasets, validation of large language model-as-judge, and metric design.
  • Experience developing applications or infrastructure using large language models and AI agents, including function calling, Model Context Protocol, or multi-agent systems.
  • Experience providing and promoting internal platforms and developer tools through platform engineering or DevOps.
  • Experience at an artificial intelligence or machine learning startup.
  • Experience with data governance and compliance in large-scale organizations.
  • A track record of open-source contributions or technical advocacy.
  • A Bachelor's or Master's degree in Computer Science or a related field.

About the Company

Sakana AI

Orchestrates multiple LLMs via API

Sakana AI builds AI foundation models and a multi-agent orchestration platform. Its flagship product, Sakana Fugu, runs as a single language-model interface that plans, delegates to, and combines outputs from a swappable pool of LLMs—including closed frontier models, open models, and its own instances—behind an OpenAI-compatible API. Fugu comes in two flavors: a low-latency everyday model and Fugu Ultra for tackling tough, multi-step tasks in software engineering, scientific research, and cybersecurity. The company differentiates itself by treating orchestration as a core capability to hedge against vendor lock-in and export-control shifts, using techniques like evolutionary model merging and agentic research systems. Sakana aims to provide reliable, flexible AI tooling for developers and enterprises through its API and tool integrations, while continuing to publish research on learned model orchestration.

Company size
201-500
Company stage
Series B
Total funding
$365M
Headquarters
Tokyo, Japan
Founded
2023

Recently Posted Jobs

Explore more opportunities at Sakana AI

See more jobs