R&D Software Engineer

We are looking for a passionate developer who loves designing, building and testing every layer of the system they work on. You care about code maintainability and about continuously improving software that has already shipped. We need your help to expand and evolve our modules to make them more flexible, open and robust.
You'll see the products you work on adopted and turned into powerful systems that support the critical business needs of our customers (some of them are listed here: http://www.answermodules.com#somecustomers).
Being a full-stack developer at AnswerModules means committing to quality and performance in everything you write. That holds whether your code will be used by your colleagues, by our partners or by all of our customers: your work is going to make the difference.

How we work: AI-assisted engineering

We believe AI is changing software engineering as deeply as high-level languages once did, and we have chosen how to embrace it. We don't do "vibe coding". We practise AI-assisted engineering: AI is a powerful collaborator, but it doesn't replace engineering judgment.

• People stay in control. Engineers make the decisions and own what ships.
• Understanding comes before acceptance. If you can't explain it, it isn't done.
• Quality is non-negotiable. AI makes us faster; discipline makes us better.
• We learn together. AI should speed up learning, not replace it, and we share what we discover as a team.
• If you want it even shorter, the opening paragraph can stand alone, followed by just one line: "People stay in control, understanding comes before acceptance, and quality is non-negotiable.

• At least 3 years of professional software development experience.
• Solid backend development in Java (11 to 21). Groovy and/or Kotlin, and Spring Boot, are a plus.
• Strong JavaScript, HTML5 and CSS3, with experience building responsive web applications that work across devices and browsers.
• Experience with a modern front-end framework. We use Vue.js 3 with Vite for new UIs; React experience is welcome too.
• Comfortable in a codebase that mixes modern and older technologies (e.g. RequireJS, Handlebars, LESS, Grunt), with experience maintaining and refactoring complex web applications.
• Can test what you build, both unit and end-to-end (e.g. JUnit, Vitest).
• Experience with build tools, especially Maven and npm-based front-end builds. Gradle is a plus.
• Good knowledge of web servers, REST APIs and relational databases (we use PostgreSQL).
• Hands-on experience building with LLMs (model APIs, prompt and context design, RAG, tool calling/MCP), and everyday use of AI coding assistants with a clear view of where they fail.
• Familiar with agile methodologies, and able to handle fast-changing schedules and priorities.
• BS in Computer Science or equivalent experience, and a passion for web technology.

Nice to have: experience with OpenText Content Server or other ECM/EIM platforms; Docker and Kubernetes; document processing (PDF/Office generation, conversion, OCR, e-signature).
Your application: show us how you think

Along with your CV, please send us a short piece of work: a prompt you would give an AI coding assistant to implement a feature of your own invention around document manipulation.

1. The idea. Come up with a feature that would be genuinely useful to people who work with business documents. It could involve generating, transforming, comparing, splitting or merging, redacting, or extracting information from documents, or something else entirely. The choice is yours, and originality counts.

Assume a document management platform where:

• documents live in folders and have versions, metadata and permissions;
• end users work through a web UI;
• developers extend the platform through a server-side scripting API.

Anything else you need, assume it and state your assumption.

2. The prompt. Write the instructions you would give the model to implement your feature. We want to see:

• how you describe the problem;
• what context you give the model;
• what constraints and non-goals you set;
• how you define "done" (acceptance criteria, tests);
• which edge cases and failure modes you anticipate (permissions, large files, unsupported formats, …).

3. A short note (max 300 words). Tell us who the feature is for and why it matters. Then tell us what you expect the model to get wrong, and how you would check its output.

Format: a single PDF or Markdown file, with the prompt no longer than two pages. You don't need to run the prompt or send any code: we evaluate the prompt and the thinking behind it.

What we look at: the value and originality of the idea, clarity and structure, the quality of the context and constraints, how testable your definition of "done" is, and how well you anticipate what can go wrong.

Write it yourself. From an AI-assisted team, this may sound surprising, but the point of the exercise is to see how you think, not how a model thinks. Submissions generated with AI will be disqualified without further consideration. If you're invited to an interview, we'll go through your submission together. We'll ask you to explain your choices, then change some of the requirements and see how you adapt your prompt.

At AnswerModules you will work in a fast-paced environment where your impact is visible from day one. If you jump on board, we can't promise an easy ride. We can promise one of the most rewarding experiences of your career: you'll learn a lot and test your whole skill set on multiple products that are already used by hundreds of thousands of users worldwide. You'll also be part of a team that is actively working out what professional software engineering looks like in the AI era.
We expect a lot from you, but your efforts will be rewarded by great benefits:

• Competitive salary
• Top-notch equipment
• Flexible work hours
• Smart Working
• Performance Bonus
• Training & Development
• Free chocolate
• Much more...