Internships Data Science/ai

At Agidens, we enhance process efficiency, quality and safety in sectors like life sciences, food and beverage, chemicals and infrastructure. Embracing the digital era, we have established a dedicated Data Science and Artificial Intelligence team to innovate and provide data-driven solutions. This strategic shift towards DS/AI highlights our commitment to using advanced technologies to tackle complex challenges and improve operations, helping our clients predict outcomes, make smarter decisions, and achieve superior operational efficiency.

**What will you do?**:
Topic 1 - Large Language Model Orchestrator

Problem description:
As Agidens increasingly utilizes specialized LLM bots for various operational tasks, the challenge of managing multiple bots through seperate user interfaces has emerged. This fragmentation can lead to inefficiencies and a steep learning curve for employees, hindering overall productivity.

Proposed solution:
The project will focus on creating an LLM-driven "Orchestration Bot" that serves as a central hub within Microsoft Teams, where it will intelligently route user inquiries to the appropriate LLM bot based on the question and contextual clues. This integration aims to simplify access and interaction with our suite of bots, making them more accessible and efficient for all employees.

Topic 2 - Advanced Insights for Industrial Processes

Problem description:
Proposed solution:
Topic 3 - Automated Metadata Discovery

Problem description:
Agidens implements industrial automation solutions for production companies in various industries. These systems often have 10,000s of configuration parameters such as alarm thresholds, control settings, etc. In complex organizations with legacy code bases it is very difficult to identify all potential settings as the automation systems do not have clear metadata differentiating the settings from sensor values. This makes it difficult to monitor these parameters, causing risks on the continuity and security of the installation.

Proposed solution:
Building an AI system that can infer metadata of each value (“tag”), such as the type (e.g., parameter, sensor value) or the context (e.g., “temperature” vs “analogue range” vs...), using the historical values and variation of values, various naming conventions, type of operations in the code, etc.

**Who are you?**:

- Currently enrolled in the 3rd year of a bachelor’s program or in a master's degree in computerscience, physics, mathematics, or a related field
- Strong proficiency in Python
- Excellent verbal and written communication skills in English
- Experience with deep learning libraries such a PyTorch is a plus
- Familiarity with cloud deployment is a plus
- A basic understanding of industrial processes is a plus

**What do we offer?**:
**Let's get together**:
online meeting. During this meeting, we will discuss the internship topics you find most interesting
and conduct a small technical assessment to gauge your suitability for the role. Applications are accepted until the 1st of June. We encourage early submissions, as selections will be