Beschreibung

AI Engineer for a high-growth startupJob descriptionOn behalf of our client, an innovative, high-growth company specializing in the development of SaaS solutions, we are looking for an AI Engineer to strengthen a team currently being built around artificial intelligence. In a product-oriented environment with concrete use cases, you will be in charge of designing and deploying solutions for processing documents and complex data. The role involves a strong dimension with a direct impact on tool performance and business value, particularly via unstructured data.Main missions- Build document processing pipelines (PDF, scans, emails, attachments)- Extract and structure data from unstructured sources (invoices, delivery notes)- Set up systems for classification and semantic tagging of documents- Design and implement reconciliation logic between different documents (invoices, purchase orders, deliveries, offers)- Detect anomalies (quantity discrepancies, price differences, inconsistencies, duplicates)- Structure data to enable analysis and scenario modeling- Deploy robust AI solutions in a production environment- Participate in the continuous improvement of pipelines via feedback and learning loops- Collaborate closely with technical and product teams in a horizontal structure- Contribute to the definition and execution of the AI roadmapWe are looking for a profile meeting the following criteria:- Excellent command of Python and SQL- Proven experience in building AI pipelines / in production (beyond simple prototypes)- Concrete experience with OCR systems and document processing- Good command of LLM model usage (structured outputs, chaining, error management, prompt engineering expertise)- Expertise in data engineering (ETL, data cleaning, data quality)- Experience with Google Cloud, particularly with Google Document AI- Good knowledge of cloud and DevOps tools (ideally Kubernetes, BigQuery, Cloud Functions, Vertex AI)- Ability to work with real, complex, messy, and imperfect data- Experience in probabilistic matching and confidence scoring systems- Ability to design robust systems with fallback mechanisms- Approach oriented towards continuous improvement of models and pipelines- Available quickly jid3eeef5caen jit0518aen jpiy26aen