Research Assistant Position:

This project develops culturally aware Sinhala language resources and methods for Large Language Models (LLMs). Sinhala is a low-resource language with complex grammar, script, honorifics, code-mixing, and culture-specific expressions that are often poorly handled by current AI systems. The research aims to create datasets, benchmarks, and alignment methods that help LLMs understand Sinhala culture and generate linguistically accurate, culturally appropriate, and safer responses.

Supervised by:

Dr. Nisansa de Silva

Dr. Surangika Ranathunga

We are looking for ...

... a motivated individual passionate about Natural Language Processing and low-resource languages. The candidate will contribute to developing culturally aware Sinhala LLMs through facilitating dataset creation, analysis, and model evaluation. Strong problem-solving skills, curiosity, and the ability to work independently and collaboratively are essential.

Qualifications Required:

• An undergraduate degree in Computer Science, Data Science, or a related field with a 2nd-class upper division or above.
• Strong programming skills (Python preferred) and solid problem-solving ability.
• Familiarity with Machine Learning and Natural Language Processing concepts.
• Experience with deep learning frameworks (e.g., PyTorch, TensorFlow) is an advantage.
• Interest in low-resource languages, linguistics, or cultural aspects of AI is beneficial.

What We Offer:

• An attractive monthly stipend.
• The selected candidate(s) will be registered to pursue an MSc (by Research) at the Department of CSE, UoM.
• Up to 60% tuition waivers on the degree pursued.
• Opportunity to contribute to impactful research publications.
• Support for advancing your research career.

Responsibilities:

• Coordinate the process of Collecting and annotating culturally relevant datasets for NLP tasks..
• Evaluate Large Language Models for linguistic accuracy and cultural appropriateness.
• Analyse model outputs to identify biases, errors, and cultural misalignments.
• Assist in developing and fine-tuning models for improved cultural alignment.
• Document findings and contribute to research publications.

Academic Expectations:

• Please read the FGS MSc page to learn about university academic expectations.
• Please read the MSc (Research Degree) parts of this page to learn about Dr de Silva's research supervison expectations.

How to Apply: