About this role
Exelab · R&D · Full-time · Remote, Italy · Reports to the R&D Lead The R&D team at Exelab Exelab builds custom, AI-powered solutions around real business needs. Our R&D team explores new technical approaches, develops reusable software and turns useful experiments into capabilities that the wider company can put to work. We are adding a mid-level engineer to strengthen this team, working alongside the R&D Lead across Polyant and PlugSync. Polyant is our runtime for AI agents; PlugSync is our integration and data synchronization platform. You will contribute to both products, with priorities and focus agreed with the R&D Lead. The role You will take defined problems from investigation through implementation, testing and documentation. Some work will start as an experiment or proof of concept; other work will improve an existing feature, integration or operational workflow. You will help decide when an experiment is ready to become supported software and what is needed to make that transition safely. This is a mid-level role for someone who has already delivered and maintained software in a professional setting. You should be comfortable making progress independently on a scoped piece of work, proposing an approach and raising uncertainties early. The R&D Lead will set priorities with you, provide technical guidance and review consequential architectural decisions. You will work across Polyant and PlugSync over time, rather than splitting every day between them. The aim is sustained progress on agreed outcomes, with clear ownership of individual features or experiments. What you will do Build and improve scoped features, integrations and internal tools across Polyant and PlugSync. Investigate technical questions, build small prototypes and evaluate their usefulness before expanding them. Contribute to AI agent capabilities and workflows in Polyant, including LLM APIs, tool use, retrieval and evaluation. Develop and maintain API integrations and synchronization workflows in PlugSync, considering authentication, data mapping, retries, duplicate processing and error recovery. Turn validated prototypes into maintainable software with tests, documentation, observability and a clear deployment path. Use AI coding tools as part of everyday engineering, while checking correctness, security and maintainability. Contribute reusable CLI tools, skills, templates and playbooks when they address a demonstrated need in the product or delivery teams. Debug defects and operational issues in your areas, communicate what you find and improve the system to prevent recurrence. Share findings through concise documentation, demos and collaboration with delivery engineers when an R&D result is ready for adoption.