Data scientist and application developer
Pubblicato il 01-08-2026 - Johnson & Johnson in Italia
Job Function: R&D; Product Development Job Sub Function: R&D; Digital Job Category: Scientific/Technology All Job Posting Locations: Latina, Italy Job Description: Stiamo cercando il miglior talento in ruolo di Data Scientist, Application Developer. Questo ruolo è disponibile in più paesi e può essere pubblicato con numeri di requisition diversi per soddisfare i requisiti normativi locali. Concentriamoci sui paesi in linea con le tue aspettative e preferenze: United States – R-079813; Italy – R-082162. Le tue applicazioni saranno considerate come una unica applicazione. Scopo: We are seeking a Data Scientist and Application Developer to join a cross‐functional team of engineers and scientists delivering data‐driven solutions tied to manufacturing. In this role you will alternate between building business applications that enable fast, iterative analytics and operationalizing models that directly inform product quality decisions in a GMP context. Responsabilità: Build and maintain internal web applications that make analytics usable in day‐to‐day workflows.
Develop services using Python with Flask (and/or Fast API where appropriate), including: Lightweight APIs for data access and model execution; User‐facing interfaces using Flask + Jinja, with minimal Java Script and optionally React for richer interactions. Implement pragmatic software engineering practices (modular design, testing, logging/monitoring hooks, clear documentation) aligned to intended use (GMP vs non‐GMP). Apply strong fundamentals in statistics, data extraction & cleaning, and modeling to diverse manufacturing datasets (e.g., batch/process time series, quality data, laboratory and instrument outputs). Develop, evaluate, and document models intended to support insights into manufacturing performance in a GMP environment, focusing on interpretability and lifecycle considerations. Collaborate with engineers, scientists, and quality partners to define the problem, select fit‐for‐purpose methods, and trans