Topic

AI Mapping Service for APIs

Motivation and goals

APIs play a central role in the integration of applications. Interface descriptions enable the automatic generation of calls. Different interface descriptions can be mapped to each other. E.g. they can have the same elements but with different identifiers (e.g. street or street). Transformer models enable various NLP tasks like translations, summaries, query rewriting or code completion. With an AI mapping the same call should be used for different interfaces. The goal of this work is to develop a mapping service for interface descriptions.

Tasks

• Investigation and evaluation of the transformer technology for mapping or modification of interface descriptions
• Development and implementation of the AI Mapper and a service that modifies the call and serves different interfaces
• Build a corpus of different interface descriptions and demonstrate the system in appropriate scenarios 
• Develop criteria for evaluating the system

An AI computer with 4 Nvidia 3090 GPUs can be used for the work.

Location: Karlsruhe and/or Bretten

Contact Recruiting:

Daniel Iwtschenko

+49 7252 96-2224

LinkedIn - XING