Master Thesis Project - Sinch Sweden - AI Center of Excellence

Master Thesis Project – Sinch Sweden – AI Center of Excellence

Sinch is a global leader in cloud communications, helping businesses connect with their customers on their mobile phones. We reach every phone on earth, with over 147 billion conversations every year. Our technology powers the world's leading communications platforms. 

Considering the current state of art in Generative AI, we are interested in projects trying to understand how Large Language Models can be used in novel ways on Natural Language Processing tasks. We are also curious about the relationship between NLP and the new multimodal approaches. 

We are looking for Master Thesis students in Computer Science, Machine Learning, Artificial intelligence, and related fields that want to develop their thesis project on topics related to Generative AI. 

Information about the Master Thesis Project: 

• Duration: minimum 3 months up to 6 months.
• Partnership with another student is accepted.
• Themes:

- LLMs & Prompt Engineering: Techniques for adjusting prompts, understanding their impact on the generated output, and best practices for optimizing language models. 

- LLM Evaluation: Methods to assess the performance of language models, including measuring response accuracy for specific tasks, ensuring stability, identifying hallucinatory responses, and defining relevant metrics. 

- LLM and RAG: Determining the best approach to apply RAG techniques (Retrieval Augmented Generation) and how it compares with alternative solutions like fine tuning, scaling models, etc. 

- Multi-modal LLMs: Exploring the practical functioning and expected performance of models that process multiple data types, identifying the tasks these models excel at. I.e. How good are they at extracting information from images using prompts.

- Prompt-Based Interfaces: Creating software products centered around prompt-based interactions, understanding the requirements for providing prompt-driven services.

Please do mention in your application what theme you would like to develop your Master Thesis Project.