Forward Deployed Engineer - Japan

Cursor; Singapore; Remote; SGD 120,000 - 180,000; Full time

Job summary

Cursor is seeking an experienced software/ML engineer to embed with customer teams and own an outcome-driven, production-grade SpaceXAI system. You will ship agents and AI workflows, wire them into customer auth, data, and tools, and ensure operability after departure.

You will lead delivery from discovery through launch, iteration, and post-launch support, owning systems end-to-end in real-world environments. Fluency in Japanese and English is required.

Qualifications

• 5+ years of experience in software/ML engineering or data science.
• Proficient in production code (Python, JavaScript/TypeScript).
• Owned a customer or operator outcome, shipping end-to-end systems.
• Built AI-native workflows or agents in production and debugged real failures.
• Experience with production reliability: metrics, alerts, safe rollouts, incident response.
• Ability to build end-to-end: frontend, backend, infra, and prompts.

Responsibilities

• Lead discovery with customers, identify bottlenecks, and define success metrics.
• Design and ship production agents, AI applications, and coding-agent workflows on Grok and models.
• Get a first version live in days and harden it through rollout and iteration.
• Own production quality: tracing, evals, debugging, latency, and cost tradeoffs.
• Build systems around the model and integrate with customer auth, data, and delivery systems.
• Measure outcomes in terms of revenue, cost savings, time, or defects.
• Collaborate with Staff+, platform, and domain leaders; communicate tradeoffs clearly.
• Turn learnings into reusable patterns for SpaceXAI products.

Skills

Python

JavaScript/TypeScript

Machine learning

Data science

End-to-end development

Ambiguity handling

Customer outcomes

System integration

Tools

CI/CD

Retrieval systems

Job description

Our mission is to automate coding. The first step in our journey is to build the best tool for professional programmers, using a combination of inventive research, design, and engineering. Our organization is very flat, and our team is small and talent dense. We particularly like people who are truth-seeking, passionate, and creative. We enjoy spirited debate, crazy ideas, and shipping code.

About the role

You embed directly with customer teams as the engineer on the ground: own an outcome, ship a production-grade SpaceXAI system that still runs after you leave, and become the critical technical feedback loop from the field back to product engineering.

The work is the project. One week you might take a messy business workflow and put a production agent on it: pick the use case, stand up the architecture, wire it into the customer’s auth, data, and tools, then harden it with evals until the team actually uses it. Another week you might live in a customer’s repo and delivery system and ship an agent-powered path for a migration, a PR review loop, or an incident-to-fix flow, behind tests, review, and rollout rather than in a chat sidebar. Either way, you measure what matters, leave the customer able to operate it, and bring the pattern back as another use case for SpaceXAI.

This is not a demo role. You own the work end to end, from the first discovery call through launch, iteration, and post-launch support, and you are responsible for systems that work in the real world.

What you’ll do

• Lead discovery with the customer: orient in their codebases and workflows, find the real bottleneck and the use case worth building, and define clear success metrics

• Design and ship production agents, AI applications, and coding-agent workflows on Grok and our models, on live business processes and inside the customer’s engineering org

• Get a first version live in days, then harden it over weeks with rollout, monitoring, and iteration from real usage

• Own production quality: tracing, evals, debugging model or delivery failures, and latency and cost tradeoffs

• Build the systems around the model (tools, MCP servers, retrieval, rules, skills, CI gates, and evals) and integrate them into the customer’s auth, data, workflow, and delivery systems; decide when prompting, architecture, data, process, or a model change is the right lever

• Measure the outcome that matters (revenue, cost, hours saved, error rate, cycle time, escaped defects), not seat adoption

• Work directly with Staff+, platform, and domain leaders, going deep in their systems while communicating clearly about tradeoffs and results

• Leave the customer team able to run what you built, and turn what you learn into reusable patterns and improvements to SpaceXAI products

You may be a fit if

Required

• You have 5+ years of experience in software engineering, machine learning engineering, or data science

• You write and review production code (Python, JavaScript/TypeScript; other languages welcome)

• You’ve owned a customer or operator outcome, not just a slide or a prototype, and can turn a fuzzy problem into a scoped, shipped system

• You’ve built and owned AI-native workflows or agents in production, and debugged real production failures (model, tool, data, pipeline, or delivery path)

• You’ve handled production reliability: metrics, alerts, safe rollouts, incident response

• You build end to end: frontend, backend, infra, and prompt iteration

• You thrive in ambiguity and don’t need a complete spec before you start building

• You can talk to working engineers and to a VP

• Fluency in both Japanese and English

Nice to have

• 2+ years in a customer-facing role, leading discovery conversations and being accountable for outcomes for external stakeholders

• Hands-on depth in several of: agents and tool-calling, eval harnesses, retrieval and context systems, CI/CD or developer platforms, and AI-native coding workflows in a real repo

• Experience as a forward-deployed, solutions, or platform engineer embedded with customers, or as a founding engineer who owned an AI product end to end