Thesis Projects at HMS LABs

A SURVEY ON THE STATE OF THE ART IN LARGE LANGUAGE MODELS AND GPT ALTERNATIVES

Department: HMS Labs
Size: 1-4 students
Scope: Artificial Intelligence, Large Language Models (LLMs)

Description:
The interest in AI is higher than ever. The release of ChatGPT by OpenAI in 2022 has sparked a new AI craze. There are a multitude of startups making new services that are based on OpenAIs APIs. And everyone wants to get in on the trend. OpenAI is the current market leader in providing Large Language Models as a service and tech giants like Google, Microsoft, and Facebook are struggling to keep up in spite of their best efforts.

This thesis project focuses on making a survey on the state of the art in Large Language Models, who the providers are, what open source alternatives are available, and how they compare to each other.

Prerequisites:
Bachelor in computer science or similar. Experience in a programming environment suitable for AI development like Python or Matlab, knowledge in artificial intelligence and LLMs, and experience in C/C++, make, and CMake is an advantage.

DESIGN OF A RASPBERRY PI BASED INDUSTRIAL GATEWAY

Department: HMS Labs / Anybus
Size: 1-4 students
Scope: Electronics, Embedded Design, Hardware Design

Description: Since their first release, Raspberry Pi (RPI) single-board computers have become incredibly popular among home automation enthusiasts, computer and electronic hobbyists, etc. Their popularity among the general public has led to an increasing popularity among industrial customers. Today, the Raspberry Pi Foundation sells 70% of all the Raspberry Pi made to industrial customers. The RPI hardware community and ecosystem potentially have many benefits for industrial use cases.

The task is to develop a hardware design for a carrier board for an industrial Raspberry PI module that fits inside an existing housing, with some modifications.

The projects include designing and drawing an industrial PCB carrier board that eventually could be produced in our factory for further investigation.

Prerequisites:
Bachelor program within embedded electronics or similar. Experience in electronics and the use of PCB design software is an advantage.

ANALYZE AND COMPARE METHODS OF BRIDGING LAYER 2 INDUSTRIAL ETHERNET OVER IP-BASED NETWORKS

Department: HMS Labs / Anybus
Size: 1-2 students
Scope: Industrial Communication, Communication Analysis, 5G, Ethernet

Description: Data exchange on wireless telecom networks (such as 4G and 5G) is based on IP communication, enabling protocols such as TCP/IP and UDP/IP messages to be transported over the wireless communication link. However, many industrial protocols are not using IP as the base for the communication and instead use other layer-2 communication.

There are new methods being added to the 3GPP cellular communication standard (such as Ethernet PDU's and TSN), which will make it possible to transport layer-2 messages directly over the cellular network. There are also methods (such as GRE, VXLAN, L2DP) that can be implemented to tunnel the layer-2 communication over the IP-based cellular network.

The goal with this thesis project is to analyze and compare the different methods to bridge layer-2 communication over the wireless cellular network and consider performance, cost, ease of implementation, availability and other things of importance related to the different methods.

Prerequisites:
Bachelor program within embedded electronics or data science. Experience in electronics and the use of PCB design software is an advantage.

ANOMALY DETECTION ON INDUSTRIAL SENSOR DATA

Department: HMS Labs
Size: 1-4 students
Scope: Anomaly detections, Predictive Maintance, Artificial Intelligence

Description: There is an increasing interest in intelligent applications for industrial use cases. Topics of particular interest are predictive maintenance and anomaly detection. One of the main challenges is to lower the threshold enough to enable users with a non-data science background to be able to reek the benefits of artificial intelligence (AI).

One common use case in industrial manufacturing where AI can be applied is anomaly detection. There are thousands of devices on a factory floor generating time series data at a high frequency. Traditionally, these signals are monitored manually and through fixed thresholds. While thresholds are a robust way of ensuring signals stay within specified tolerances, they cannot be used to find more complex correlations.

This project aims to develop an anomaly detection system that can operate on one-dimensional signals and find anomalous states. The system should train multiple models on a given signal, find the best one, and then deploy it in an "operational" setting. The project and scope can be adjusted to the size of the group and the participants.

Prerequisites:
Bachelor's or master's program in data science, computer science, or similar. Experience in a programming environment suitable for AI development like Python or Matlab, knowledge in artificial intelligence and machine learning, and experience in C/C++, make, and CMake is an advantage.

ABOUT US

At HMS, we believe that connecting matters in everything from making machines talk to engaging with our customers. Our state-of-the-art technology provides a fast and easy way to network millions of industrial devices all over the world. From factory automation to connected buildings, remote solutions, and future technologies, we provide innovative solutions that save you time and money creating a more productive and sustainable world. HMS stands for Hardware Meets Software ™.

HMS is a truly international company, and we like to think that we share a common culture. Heart, Mind and Soul is a summary of how we do things (makes for a pretty good abbreviation too).

Welcome to HMS, this is HMS!

ABOUT THE EMPLOYMENT AND THE APPLICATION

 If you want to know more about the recruitment process, please contact HR Andrea Wagner, anwa@hms.se.

Welcome with your application!