Master Thesis : AI-Powered Troubleshooting Agent

Description

At ELCA, we support and maintain a wide range of enterprise solutions based on Microsoft Dynamics 365. These solutions involve custom developments using C#, JavaScript and more, tracked through Azure Dev Ops (ADO) using work items (User Stories, Tasks, Bugs) and source code repositories (Git). During maintenance phase, customer reports issues (bugs and support need) by submitting tickets via ADO (Azure Dev Ops).

Currently, our consultants manually analyze these tickets, make sure they have enough context to reproduce the case and base their analysis with historical knowledge, then inspect related code to diagnose and resolve bugs. This approach is time-consuming, potentially requires back-and-forth with customers and depends heavily on human expertise and familiarity with past project history.

We are in the AI era; in this internship we propose to leverage machine learning and language models to enhance software maintenance and support. By building an AI-powered troubleshooting agent that has access to the code and documentation and is able to learn from past project history, bug reports, and associated code changes, we can drastically reduce resolution times, improve accuracy, and empower support teams with contextual suggestions.

This AI solution will not only streamline operations but also contribute to building a dynamic and self-improving knowledge base over time.

As an intern, you will join a team of experts and passionate professionals working at the crossroads of enterprise IT and artificial intelligence. You will evolve in a stimulating, collaborative environment that encourages innovation, initiative, and technical excellence. You’ll be supported by experienced mentors, have access to modern tools and infrastructure, and contribute to a project with real impact for customers and internal teams alike.

The objective of this internship is to design and build an AI-based troubleshooting assistant that will:

• Crawl and index historical ADO tickets
• Link ADO tickets to code segments
• Use NLP techniques and vector-based semantic search
• Optionally:
  Integrate code understanding via models like Qwen2/Codellama or other to suggest code changes

• Build a lightweight web-based UI in the language of your choice.
• Authentication + access to user-specific ADO projects via Microsoft Graph API or Azure Dev Ops API.

• A dynamic work and collaborative environment with a highly motivated multi-cultural and international sites team
• The chance to make a difference in peoples’ life by building innovative solutions
• Various internal coding events (Hackathon, Brownbags), see our technical blog
• Monthly After-Works organized per locations

• Strong skills on AI & NLP/ML.
• Programming:
  Strong skills on object-oriented programming, ideally C#. And skills with Java Script.
• Experience with LLMs (OpenAI, LLama, Qwen2, etc.).
• Dev Ops concepts.
• Databases: SQL or No
  
  SQL
• Version Control:
  Git

• Knowing Azure Dev Ops (ADO).
• Vector search tools (e.g., FAISS, Azure AI Search, Pinecone)
• Web Dev:
  React.js, Blazor, or ASP.NET Core
• Azure services:
  Functions, Cognitive Search, Azure OpenAI
• Familiarity with Dynamics 365 project structures.

This internship starts in February 2026
.

Applications must include your most recent academic transcripts (grades); applications without transcripts will not be considered.