Development and Evaluation of a Lightweight Chatbot for Technical Troubleshooting of Munters Dehumidifiers

Background:
Industrial devices and systems often require troubleshooting support to resolve operational issues. However, complex user interfaces/display and limited access to real-time assistance can hinder effective problem resolution. An offline chatbot, designed for lightweight deployment on edge devices, offers a promising solution by providing an accessible, standalone tool for guiding users through troubleshooting without requiring internet connectivity.

This thesis focus is developing a lightweight chatbot solution tailored for small-scale troubleshooting scenarios (e.g., for industrial devices like Munters dehumidifiers). The project will explore the design, implementation, and evaluation of the chatbot, emphasizing efficiency, usability, and edge device compatibility.

Allowing for a more comprehensive development process, this thesis will possibly be conducted in collaboration with another student who will focus on UI/UX area.

Objective:
The primary objective of this thesis is to design, implement, and evaluate a lightweight chatbot solution (as proof of concept) that:

Enhance the user interaction of our dehumidification products during alarms, warnings, and observations Provides accurate and user-friendly troubleshooting assistance. Is optimized for deployment on limited computational resources.

Scope of Work:

  • Investigate existing chatbot technologies and rule-based systems.
  • Review lightweight natural language processing (NLP) approaches suitable for edge devices.
  • Develop the architecture for the chatbot.
  • Design logic of the engine for mapping user inputs to predefined troubleshooting steps.
  • Create an FAQ database containing structured questions, keywords, and responses.
  • Develop a chatbot prototype using lightweight frameworks (e.g., Python, SQLite for database management).
  • Integrate basic NLP techniques to improve input recognition and flexibility (e.g., keyword matching, intent classification).
  • Optimize the chatbot for low-resource devices by minimizing memory and computational requirements.
  • Test the chatbot with a simple example of real-world troubleshooting scenarios.
  • Evaluate performance metrics such as response accuracy, resource consumption, and user satisfaction.

Skills Required:

  • Academic Background: Computer Science/Software Engineering or Artificial Intelligence/Machine Learning
  • Programming skills (e.g., Python or a similar language).
  • Familiarity with basic NLP techniques and rule-based systems.
  • Familiarity with database
  • Analytical skills for performance evaluation and user testing.

Application:

We will review the applications as they come in, so please send your application as soon as possible but no later than Jan 12th.

Contact:

Zeinab Moradi Nour, R&D manager AI & control

zeinab.moradinour@munters.com

Location

Kista (MEA)

Job Overview
Job Posted:
2 weeks ago
Job Expires:
Job Type
Full Time

Share This Job: