Software Developer, circular-construction AI (REON / SXL)

ppbbMission /b /b /p /ppThe Structural Xploration Lab (SXL) at EPFL is developing REON, an AI system that predicts the reuse value of building components before demolition, turning research validated on 23,200 real reuse-market transactions into a working product. We are looking for a software developer to help build the core of this product alongside the founder: taking a validated research prototype and turning it into a clean, working, deployable tool. /pppbbMain duties and responsibilities /b /b /p /p liExtract and harden a scoring engine from an existing Python/Streamlit prototype into robust, standalone code /li liRebuild the data layer from Neo4j to a Python-native graph structure /li liImplement data-ingestion and similarity-matching logic that completes missing component information (GNN) /li liBuild a lightweight web interface: data upload in, ranked results out /li liWork independently from rough goals to working software, using modern tooling (including AI coding assistants) to move fast /li liKeep the codebase clean and documented so it can be extended by others. /li ppbbProfile /b /b /p /p pWe are looking for a strong early-career Python developer who is interested in bringing a research prototype into a working product. The role combines graph data, ML, and practical web development. Some technologies may be learned during the project, but solid Python skills and hands-on experience with graphs and GNNs are essential. /p liDegree in computer science, software engineering, data science, machine learning, or an equivalent field; recent graduates are welcome /li liStrong Python skills and the ability to write clean, modular, tested, and documented code /li liHands-on experience with graph data structures and graph neural networks, gained through academic research, internships, professional work, or substantial personal projects /li liExperience with machine-learning pipelines, including data preparation, model training or inference, and evaluation /li liInterest in migrating an existing Neo4j model to a custom Python-native graph implementation /li liExperience building at least one complete application beyond notebooks, including a backend API and a functional web interface /li liComfortable learning unfamiliar technologies and improving an existing prototype step by step /li liFamiliar with Git; experience with LLM workflows, Docker, or cloud deployment is an advantage /li liSelf-directed, pragmatic, and motivated by research-to-product development /li liValid Swiss work authorisation; bholders of an S permit are welcome if authorised to work /b /li liFluent English; French or German is an advantage /li ppbbWe offer /b /b /p /p liThe chance to build the technical core of an EPFL spin-out in formation, in circular construction and AI /li liDirect collaboration with the founder and real ownership of what you build /li liA flexible, hands-on role with genuine impact /li liFor the right person, a possible path to a founding technical role at REON /li libThe starting date is flexible /band can be agreed upon with the selected candidate /li ppbbInformations /b /b /p /p pActivity Rate Min : 20.00 /p pActivity Rate Max : 80.00 /p pContract Type: CDD /p pDuration: 2-4 months /p pReference: 2351 /p