Internship Master: Developing a retrieval-augmented...

Rapid advancements in deep learning technology have led to the rise of Large Language Models (LLMs), which are capable of generating context-aware, human-like responses to queries. While many of such models are publicly available and perform great on general tasks, they often fall short in domain-specific tasks, making them less effective for specialized industries. Developing a domain-specific LLMs from scratch is prohibitively expensive for the majority of companies, which is why a technique called ‘Retrieval Augmented Generation’ (RAG) has recently gained traction. RAG enhances existing LLMs with data retrieval capabilities, allowing them to pull from sources like technical manuals, internal documentation or even SQL databases to provide more precise, context-aware responses when interacting with a user.

At Bestmix Software, we aim to stay at the forefront of innovation. Because we always look to provide the best services for our customers, our next step is to develop our own virtual assistant to provide customer support, streamline workflows and enhance training programs. To make this vision a reality, we are looking for a motivated master’s student with a background in data science to help us build and refine a RAG-powered AI assistant.

The assignment 

What you’ll do

We are looking for an enthusiastic master student with a background in data science to help us create our own virtual assistant. In this internship, you will be part of our data team and play a critical role in shaping how AI is used to support our customers.

Your responsibilities will include

• Researching existing RAG strategies and models
• Preprocessing data to tailor the model(s) to specific tasks
• Selecting the most suitable strategies for our business needs
• Developing a working prototype of the RAG model
• Evaluating and optimizing the model’s performance

What will BESTMIX Software offer you?

This internship offers a unique opportunity to gain experience in AI development, making you stand out in the job market. By the end of this internship, you will have:

• Practical experience in Retrieval-Augmented Generation (RAG)
• Hands-on exposure to LLM fine-tuning and deployment in real-world applications
• Experience working with real datasets, preprocessing, and adapting them for AI models
• Skills in model evaluation and optimization
• Collaboration experience within a professional data science team
• A concrete AI project to showcase to future employers, demonstrating your ability to develop solutions for real business challenges
• access to data to train and fine-tune the AI assistant
• compute resources to develop and evaluate AI models
• A collaborative and innovative work environment as part of our data team

About BESTMIX Software

BESTMIX Software stands for innovation, reliability, passion, know-how and a drive for results. Based on these corporate values, BESTMIX Software has been adding value to the several nutritional industries for 50 years. At BESTMIX Software, which became part of the Info Support International Group in 2017, you arrive in an international team made up of more than 180 committed employees. Their collected knowledge and dedication translate itself over and over again into modern solutions that offer assistance in every step of the formulation and production process of food and feed products. Therefore, it is no surprise that BESTMIX' ERP solution and formulation software are well-known names in our sector.

Starting or continuing your career at BESTMIX Software means developing all your talents, even those that are still unknown to you. We appreciate any sector knowledge, but we predominantly count on your ability to function in a team and your openness to work with other cultures. With more than 500 clients in over 60 countries you become partly responsible for the global food supply of both humans and animals.

BESTMIX Software is carried out by its staff. That is exactly why flexibility and the well-being of our employees are focus points in our corporate culture. Do you still have any questions or are you simply too eager to start applying? Don’t hesitate and contact us straight away or visit our website: .