AI-First Software Developer - Softeon

Job Description

Ship more of what customers need, faster, and prove every change works.

Our applications run the warehouses that get orders out the door on time. We are changing how that software gets built: AI agents and AI-assisted CI/CD help us deliver more functionality to our customers, faster, with higher quality.

We are looking for an AI-first developer: someone who treats AI agents as a team to direct, not a fancy autocomplete. You will pair deep Java engineering with hands-on experience building through agents to deliver great new functionality across our applications, and you will help set the standard for how the rest of engineering works this way. You know that speed only counts if it comes with proof, so automated testing, validation and agent-assisted review are part of how you ship every change.

What you will do

• Deliver great product functionality. Design and build new features and enhancements across our Java applications (receiving, inventory, picking, replenishment, shipping, integrations and user experience) that customers value and rely on.
• Build with agents every day. Use Claude Code and similar coding agents to plan, implement, refactor and test features. Break work into clear specs, give agents strong context, and know when to delegate and when to step in.
• Author and maintain agent skills. Write and refine skills, prompts, tool definitions and MCP servers that encode Softeon's domain knowledge, such as Elastic UI form and flow authoring, WMS configuration and integration patterns.
• Own a fully automated delivery pipeline. Build and run CI/CD where every change passes automated builds, unit and integration tests, static analysis, security scanning and AI agent code review before a human approves it.
• Make quality provable. Build and extend automated validation (unit, API and Playwright end-to-end tests) so that AI-generated code, configuration and integrations are verified automatically, with evidence.
• Operate what you ship. Build the telemetry, logging and monitoring your features need, and treat performance and reliability as product requirements rather than a late pass.
• Modernize with confidence. Use agents to understand, refactor and modernize existing code, add missing test coverage, and pay down technical debt without slowing feature delivery.
• Measure velocity and quality. Track cycle time, defect rates and escaped bugs, and use the data to keep improving how agents and pipelines are used.
• Work with product and customers. Turn real warehouse and supply chain needs into well-specified, testable software, and talk with product managers, customers and Ops teams in their language.
• Raise the bar for the team. Share patterns, review others' agent workflows, measure what works, and help colleagues become effective AI-first developers.

Why it is interesting

• Agents can write code faster than any team can review it. The hard part is no longer typing; it is deciding what to build, giving agents the right context, and proving that what comes back is correct in a domain where a plausible answer can still stop a warehouse. That means designing the specs, skills, tests and pipeline gates that let a small team ship enterprise software at a pace that used to need a much larger one, and there is no settled playbook to copy.
• You will do this on mature, mission-critical applications with real customers, real data and real operational stakes, which is a harder and more rewarding problem than a greenfield demo.

Frameworks and tools you will work with:

Languages & core

• Java 17+, Spring / Spring Boot, REST and messaging APIs, SQL and relational databases

Front end

• TypeScript, Angular, and Elastic UI, our metadata-driven UI platform (XSLT forms and flows rendered as JSON schema)

Scripting & extension

• Groovy scripts and connectors, extension APIs and scaffolding for customer and partner modifications

AI & agents

• Claude Code and other coding agents, agent skills, Model Context Protocol (MCP) servers, LLM APIs, evaluation of agent output

Quality

• JUnit, API and integration testing, Playwright end-to-end tests, test data and fixtures, mutation testing

Delivery & DevSecOps

• GitHub and GitHub Actions (or Jenkins / Azure DevOps), automated agent code review, static analysis, security scanning, SBOM, Docker, Kubernetes, infrastructure as code

Operations

• Logging, telemetry, monitoring and observability tooling across cloud environments

Qualifications

Required

• 2+ years of hands-on AI and agent-based development: building production software with coding agents such as Claude Code, Cursor, Copilot agents or similar, and/or building LLM-powered features and agents (tool use, retrieval, MCP, evaluation).
• Expert-level Java: modern Java (17+), Spring/Spring Boot, REST APIs, multithreading, performance tuning and clean, maintainable design.
• 5+ years of professional software development, with a track record of building, shipping and operating production systems.
• Strong experience in supply chain, automation, MRP or ERP systems. You understand how orders, inventory, materials and fulfillment flow through enterprise systems, and you can spot an answer that looks plausible but is operationally wrong.
• CI/CD and DevOps fluency: Git-based workflows, pipelines (GitHub Actions, Jenkins, Azure DevOps or similar), containers (Docker, Kubernetes) and infrastructure as code.
• Automated testing, observability and continuous delivery as part of how you build, not a separate phase.
• Prompt and context engineering skills: clear written specs, structured prompts and the judgment to evaluate and correct agent output.
• Comfortable across the boundary: you can write the front-end, service and database code a feature needs end to end.
• Solid fundamentals in relational databases and SQL, system design, and secure coding practices.
• Permanent resident of the United States, legally authorized to work in the US without the need for current or future visa sponsorship.

Also valuable

• A platform where behavior is configured through data or metadata rather than written in code, such as low-code, rules engines or schema-driven UI.
• Modern TypeScript and Angular (or React), including component architecture and design-system work.
• Direct experience with warehouse management (WMS), warehouse execution (WES), transportation management (TMS) or MRP/ERP platforms, as a developer, implementer or operator.
• Experience with warehouse automation: conveyors, sortation, pick-to-light, AMRs, goods-to-person systems or their control software.
• Mobile or progressive web applications, including RF/handheld devices, offline behavior and poor connectivity.
• Experience building MCP servers, custom agent tooling or AI-powered developer workflows.
• Familiarity with XSLT and Groovy, and cloud platforms (AWS, Azure or GCP).
• Based in the US Eastern or Central time zone.

How we work

• AI-first by default. Agents are part of the team; you are responsible for their output.
• No change ships without proof. Automated tests, agent review and validation evidence on every pull request.
• Close to customers. We learn from warehouse floors and Ops leaders, not just tickets.
• Small, senior and fast. High ownership, low ceremony, measurable results.

What success looks like

• First 30 days: you ship production changes through the full agent-assisted pipeline and know our Java platform and WMS domain.
• First 90 days: you own a feature area end to end, have contributed new agent skills or pipeline checks, and have measurably sped up delivery without lowering quality.
• First year: you have delivered significant product functionality and are a go-to person for AI-first engineering practices across Softeon engineering.

Location and eligibility

…