锦鲤求职岗位描述
About Rakuten
Rakuten group has almost 100 million customers based in Japan and 1 billion globally as well, providing more than 70 services in a variety such as e-commerce, payment services, financial services, telecommunication, media, sports, etc.
Division Introduction
AI & Data Division (AIDD) spearheads data science & AI initiatives by leveraging data from Rakuten Group. We build a platform for large-scale field experimentations using cutting-edge technologies to provide critical insights that enable faster and better and faster contribution for our business. Our division boasts an international culture created by talented employees from around the world. Following the strategic vision "Rakuten as a data-driven membership company", AIDD is expanding its data & AI related activities across multiple Rakuten Group companies.
About the Role
As a Coding Harness Engineer, you will focus on building and integrating an AI-first coding harness that brings Rakuten's in-house LLM and internal systems directly into the developer workflow. You will deeply integrate the harness with internal code repositories, knowledge bases, ticketing systems, and review tools, ensuring Rakuten engineers get a cost-efficient, secure, and AI-first SDLC experience.
This role requires strong expertise in LLM application development, retrieval-augmented generation (RAG), agentic coding workflows, and developer tooling, with a focus on maximizing coding productivity while controlling inference cost and enforcing security boundaries.
Key Responsibilities
· Design and implement the coding harness that integrates Rakuten's in-house LLM with internal systems (Git, code search, wikis, ticketing, CI/CD) to deliver an AI-first SDLC experience for internal developers.
· Build and optimize retrieval-augmented generation (RAG) pipelines over internal codebases and documentation to provide accurate, context-aware code completion, review, and Q&A.
· Implement agentic coding workflows (e.g., code generation, fix suggestion, test generation, PR review automation) that operate safely within Rakuten's internal repositories and access-control boundaries.
· Develop cost-control mechanisms (e.g., model routing, prompt caching, request throttling, usage metering) to keep LLM inference cost-efficient at organization scale.
· Enforce security and governance across the harness — code privacy, secret redaction, access scoping, audit logging, and policy-as-code guardrails.
· Integrate the harness into developer surfaces (IDE plugins, CLI, PR bots, internal portals) to maximize adoption and seamless workflow integration.
· Collaborate with the in-house LLM platform team to align harness capabilities with model capabilities (fine-tuning, function calling, long-context, structured output).
· Develop benchmarking and evaluation tooling to measure harness quality (accuracy, latency, cost, developer satisfaction) and drive continuous improvement.
· Research and apply cutting-edge techniques (e.g., speculative decoding, tool-use agents, MCP, context compression) to optimize harness performance and UX.
Mandatory Qualifications
· 3+ years of hands-on experience in software engineering, with a focus on LLM application development, developer tooling, or platform engineering.
· Deep expertise in building LLM-powered applications, including prompt engineering, RAG, function calling, and agentic workflows.
· Strong knowledge of Python (or Go) and experience building production backend services and APIs.
· Experience integrating with developer tools and systems (Git, GitHub/GitLab APIs, IDE extensions, CI/CD pipelines).
· Bachelor's or higher degree in Computer Science, Engineering, or related field.
Nice-to-Have Skills
· Proficiency in LLM frameworks (LangChain, LlamaIndex, vLLM, LiteLLM) and retrieval systems (vector databases, BM25, reranking).
· Experience with model routing, prompt caching, and inference cost optimization at scale.
· Familiarity with agentic coding tools and protocols (e.g., Model Context Protocol (MCP), OpenAI function calling, tool-use agents).
· Experience building IDE extensions or editor integrations (VS Code extension API, LSP).
· Experience with evaluation and observability for LLM apps (Langfuse, tracing, prompt regression testing).
· Contributions to open-source AI coding tools (e.g., Continue, Aider, OpenHands, Copilot alternatives).
· Experience with security and access-control design for internal developer platforms (RBAC, secret management, audit logging).
· Familiarity with Kubernetes and cloud-native deployment of LLM serving workloads.
· Experience with RAG over large-scale monorepos and enterprise knowledge bases.
Why Join Us?
· Work on cutting-edge AI-first SDLC tooling integrated with Rakuten's in-house LLM at organization scale.
· Directly impact thousands of internal developers' productivity by delivering a secure, cost-efficient coding harness.
· Collaborate with global AI/ML and platform teams on high-impact challenges.
· Opportunity to research and implement state-of-the-art agentic coding and retrieval techniques.