Master Thesis Student - AI-Powered Knowledge and Reasoning Assistant for Railway Software Systems

Req ID:499925 

At Alstom, we understand transport networks and what moves people. From high-speed trains, metros, monorails, and trams, to turnkey systems, services, infrastructure, signalling and digital mobility, we offer our diverse customers the broadest portfolio in the industry. Every day, more than 80 000 colleagues lead the way to greener and smarter mobility worldwide, connecting cities as we reduce carbon and replace cars.

Background

Alstom is a global leader and pioneer in the railway industry, driving innovation in sustainable and intelligent mobility solutions. At the core of Alstom’s railway infrastructure lies the Train Control and Management System (TCMS) — a critical software-based system responsible for both safety-critical and operational functions onboard trains.The TCMS architecture is built upon the MITRAC platform, a software ecosystem that integrates various subsystems and devices across the train. As a key contributor to the railway community, Alstom bears the responsibility of maintaining, supporting, and continuously improving these MITRAC-based systems. This includes activities such as software maintenance, system troubleshooting, platform upgrades, and root cause analysis for reported issues.Carrying out these responsibilities demands extensive technical expertise and detailed analysis of the system’s documentation — including device-specific documents, interface control documents (ICDs), and platform-level design specifications. Developers and engineers often invest significant effort in reviewing and correlating such materials to diagnose issues and develop appropriate patches, platform releases, or engineering briefs.                                  

Problem description and goals

Motivation and Problem Statement

The growing complexity of modern railway software systems, combined with the volume of heterogeneous technical documentation, has created a strong need for intelligent tools that can support developers in efficiently understanding, retrieving, and reasoning over system information.

In pursuit of this goal, Alstom has developed a prototype AI-driven chatbot tool based on Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) technology. This system demonstrates the potential of artificial intelligence to assist engineers in issue investigation by retrieving relevant knowledge from platform documentation. However, the current version remains in a prototypical stage and lacks integration into Alstom’s operational issue-tracking and problem-resolution workflows.

Thesis Objective

The primary objective of this Master’s Thesis is to advance the capabilities of the existing AI-based tool and to transform it into a robust, domain-aware, and multi-modal system that can support real-world TCMS issue analysis and knowledge retrieval.

Research Goals and Scope

The work will involve research and development in several interrelated areas of Artificial Intelligence, Natural Language Processing (NLP), and Knowledge Engineering. The student will focus on:

• Improving the existing AI-based tool using advanced knowledge retrieval techniques.
• Developing methods to include domain-specific and external data sources in the system.
• Enabling the tool to process and understand structured data, such as logs or configuration files.
• Enhancing the system’s reasoning abilities using advanced AI reasoning models.
• Exploring ways to integrate multi-modal data such as diagrams or system models.
• Evaluate and document the improvements made to the system.

Expected Outcomes

By the conclusion of this thesis, the student is expected to deliver:

• A functional and improved version of the existing RAG-based AI assistant with enhanced retrieval accuracy and reasoning capability.
• A technical evaluation demonstrating measurable improvements in retrieval precision, response coherence, and domain understanding.
• A conceptual framework for integrating the developed system into Alstom’s issue-tracking and problem-resolution workflows.

Prerequisites:

• Master’s student in Computer Science, Software Engineering, AI, Electrical Engineering, or related field.
• Strong Python programming skills (experience with AI/ML frameworks such as PyTorch, TensorFlow, or Hugging Face preferred).
• Basic understanding of Machine Learning, NLP, and Large Language Models (LLMs).
• Familiarity with Retrieval-Augmented Generation (RAG), information retrieval, or knowledge management systems is an advantage.
• Good analytical, research, and problem-solving skills.
• Effective communication skills in English (written and verbal).
• Interest in applying AI to industrial or safety-critical systems, preferably in the railway domain.

Duration: 20 weeks

Number of students: 1

Language of thesis: English

Is Swedish a language requirement? No

Possibility to work from our office: Yes

You don’t need to be a train enthusiast to thrive with us. We guarantee that when you step onto one of our trains with your friends or family, you’ll be proud. If you’re up for the challenge, we’d love to hear from you!

Important to note

As a global business, we’re an equal-opportunity employer that celebrates diversity across the 63  countries we operate in. We’re committed to creating an inclusive workplace for everyone.