DON SANJEEWA
ALWIS

QUESTIONS WORTH INVESTIGATING

Research &
publications.

A closer look at language AI.
A wider question about who it serves.

Explore Sanjeewa Alwis’s research contributions, beginning with a co-authored journal paper on Sinhala language AI.

Read the original work, meet the collaborators and follow the evidence behind the ideas.

Explore the publication
LOW-RESOURCE LANGUAGESLARGE LANGUAGE MODELSSINHALA LANGUAGE AI

SELECTED PUBLICATION

ORIGINAL WORK · DIRECT SOURCES
IEEE ACCESS2026 / JOURNAL ARTICLE

End-to-End Adaptation of LLMs for Low-Resource Languages

The study adapts LLaMA 3.1 for Sinhala through corpus preparation, vocabulary expansion, continued pre-training and LoRA fine-tuning. It examines language-model development with limited computing resources.

Read the full paper Publisher record

AUTHORS · IN PUBLICATION ORDER

  1. P. K. Udith I. Sandaruwan
  2. Nimesh M. A. Fonseka
  3. Pamith C. Salwathura
  4. Don Sanjeewa Alwis
  5. Chathura Wanigasekara
  6. Logeeshan Velmanickam

INSIDE THE PAPER

From language data
to an adapted model.

Use the paper’s methods and evaluation sections to understand the choices behind the work.

01

Prepare the dataCorpus collection and preprocessing.

02

Adapt the vocabularySinhala-specific tokenization.

03

Train and refineContinued pre-training and instruction tuning.

04

Evaluate the resultsQuantitative and qualitative assessment.

READ, REFERENCE, REVISIT

Keep the source
close to the idea.

The full paper provides the experimental setup, findings and limitations. Use its evidence and scope when referring to this work.

Cite this publication

P. K. U. I. Sandaruwan et al., “End-to-End Adaptation of LLMs for Low-Resource Languages,” IEEE Access, vol. 14, 2026. doi: 10.1109/ACCESS.2026.3693119.

Download BibTeX
View BibTeX 
@article{Sandaruwan2026SinhalaLLM,
  author = {Sandaruwan, P. K. Udith I. and Fonseka, Nimesh M. A. and Salwathura, Pamith C. and Alwis, Don Sanjeewa and Wanigasekara, Chathura and Velmanickam, Logeeshan},
  title = {End-to-End Adaptation of {LLMs} for Low-Resource Languages},
  journal = {IEEE Access},
  volume = {14},
  year = {2026},
  doi = {10.1109/ACCESS.2026.3693119},
  url = {https://doi.org/10.1109/ACCESS.2026.3693119}
}

CONTINUE EXPLORING

A broader view
of the research.

Visit Sanjeewa’s Google Scholar profile for further publication listings and citation information.

View Google Scholar

RESEARCH COLLABORATION

Start with
a good question.

For a conversation about language AI or a specific publication, share your research question, the relevant work and the contribution you would like to explore.

Contact Sanjeewa