Beschreibung

Data Engineer – AI-based Image Recognition Platform

Stuttgart, Deutschland100% RemoteFreiberuflichStart 4/2026Dauer 6 Monate60% Auslastung

Eingestellt von

Ansprechpartner

LegendsLab Team

Projekt-ID

2979251

APIsKünstliche IntelligenzKünstliche Neurale NetzwerkeComputer VisionAutomobilindustrieMicrosoft AzureInformation EngineeringDatenqualitätPythonSQLAusbildungsaktivitätenDaten- / DatensatzprotokollierungPytorchReactJSFastapiFront End

Beschreibung

Project context:

We are looking for a highly motivated freelancer to support our team in the development and optimization of a system for AI-based identification of automotive parts in the context of circular economy / remanufacturing. The system is based on image recognition (few thousand classes) and incorporates business data / metadata to improve classification results.

The solution is shall be available online (frontend and other systems).

Description:

We are seeking a Data Engineer to support the development and optimization of a computer vision–based system for identifying automotive parts, integrating image recognition models with business data and metadata in an online platform environment.

Most important skills:

• Very strong Python skills

• PyTorch and hands-on experience in computer vision

• High level of motivation and enthusiasm

• Experience with neural network based training approaches including statistical analysis and iterative improvements

• Basic SQL skills

Optional skills:

• Experience with Python backends such as FastAPI (production-grade APIs, interfaces to frontend and external systems)

• (CNNs / ViTs, fine-tuning)

• Experience with operation of cloud based infrastructure, ideally Azure.

• React for visualization of the data/results in frontend application

Main task/activity

Data engineering with the purpose of optimization/training of a neural network based classification system.

Other tasks:

• Data cleaning and data preparation (image data and metadata), support with labeling and data quality assurance

• Training and experimentation, including statistical analysis, quality measurement, and structured error analysis

• Support with serving and integration: FastAPI endpoints, request/response handling, stability and logging

• Support in the adaptation of neural networks for image recognition (large-scale classification, long-tail scenarios)

• Integration of business data / metadata (tabular features) to improve prediction accuracy