Vision

The system can autonomously move through production areas, assess whether observed situations are safe or unsafe, and report relevant findings to the EHS officer. Autonomous navigation is already implemented in the web server. As a next step, the VLA must be trained for safety assessment tasks. We can begin with a small proof of concept, but no specific outcome can be guaranteed at this stage.

• Integration of customer-provided LLM accounts (e.g. OpenAI, cloud providers, or local deployments such as Ollama)
• No token or usage costs incurred by us; all usage is billed directly to the customer’s account
• Setup and configuration of a structured data pipeline for LLM-based processing
• Backend interface for uploading and managing documents and folder structures
• Automated ingestion and indexing of customer-provided data
• Context-aware querying based on uploaded knowledge sources
• Flexible architecture for connecting different LLM providers, depending on customer preference
• Ongoing support for integration, updates, and optimization

IMPORTANT: The quoted / charged amount for this position is just an estimation. We charge 1500EUR (NET) per day based on the real time which we have spent on the development

The system is intended to support the comparison of technical drawings with incoming insulators and to identify visible deviations or defects. This includes checking whether delivered parts match the relevant specifications and whether obvious quality issues can be detected during inspection. A proof of concept can be developed to evaluate the feasibility of automated comparison and defect detection. Due to the complexity of visual interpretation, varying part conditions, and drawing-specific tolerances, no specific outcome can be guaranteed at this stage.

• Integration of customer-provided LLM accounts (e.g. OpenAI, cloud providers, or local deployments such as Ollama)
• No token or usage costs incurred by us; all usage is billed directly to the customer’s account
• Setup and configuration of a structured data pipeline for AI-based comparison and inspection workflows
• Backend interface for uploading and managing drawings, inspection files, and folder structures
• Automated ingestion and indexing of customer-provided technical documents and image data
• Context-aware querying based on uploaded drawings, specifications, and knowledge sources
• Flexible architecture for connecting different LLM or AI providers, depending on customer preference
• Ongoing support for integration, updates, and optimization

IMPORTANT: The quoted / charged amount for this position is just an estimation. We charge 1500 EUR (NET) per day based on the actual time spent on development.