Large language model (LLM) research and development has overwhelmingly focused on the world’s major languages, leading to under-representation of low-resource languages such as Irish. This paper introduces Qomhrá, a bilingual Irish and English LLM, developed under extremely low-resource constraints. A complete pipeline is outlined spanning bilingual continued pre-training, instruction tuning, and the synthesis of human preference data for future alignment training. We focus on the lack of scalable methods to create human preference data by proposing a novel method to synthesise such data by prompting an LLM to generate "accepted" and "rejected" responses, which we validate as aligning with L1 Irish speakers.To select an LLM for synthesis, we evaluate the top closed-weight LLMs for Irish language generation performance. Gemini-2.5-Pro is ranked highest by L1 and L2 Irish-speakers, diverging from LLM-as-a-judge ratings, indicating a misalignment between current LLMs and the Irish-language community. Subsequently, we leverage Gemini-2.5-Pro to translate a large scale English-language instruction tuning dataset to Irish and to synthesise a first-of-its-kind Irish-language human preference dataset. We comprehensively evaluate Qomhrá across several benchmarks, testing translation, gender understanding, topic identification, and world knowledge; these evaluations show gains of up to 29% in Irish and 44% in English compared to the existing open-source Irish LLM baseline, UCCIX. The results of our framework provide insight and guidance to developing LLMs for both Irish and other low-resource languages.

This paper sets out the first web-based transcription system for the Irish language - Fotheidil, a system that utilises speech-related AI technologies as part of the ABAIR initiative. The system includes both off-the-shelf pre-trained voice activity detection and speaker diarisation models and models trained specifically for Irish automatic speech recognition and capitalisation and punctuation restoration. Semi-supervised learning is explored to improve the acoustic model of a modular TDNN-HMM ASR system, yielding substantial improvements for out-of-domain test sets and dialects that are underrepresented in the supervised training set. A novel approach to capitalisation and punctuation restoration involving sequence-to-sequence models is compared with the conventional approach using a classification model. Experimental results show here also substantial improvements in performance. It is intended that will be made freely available for public use, and represents an important resource researchers and others who transcribe Irish language materials. Human-corrected transcriptions will be collected and included in the training dataset as the system is used, which should lead to incremental improvements to the ASR model in a cyclical, community-driven fashion.

This paper describes ÉIST, automatic speech recogniser for Irish, developed as part of the ongoing ABAIR initiative, combining (1) acoustic models, (2) pronunciation lexicons and (3) language models into a hybrid system. A priority for now is a system that can deal with the multiple diverse native-speaker dialects. Consequently, (1) was built using predominately native-speaker speech, which included earlier recordings used for synthesis development as well as more diverse recordings obtained using the MíleGlór platform. The pronunciation variation across the dialects is a particular challenge in the development of (2) and is explored by testing both Trans-dialect and Multi-dialect letter-to-sound rules. Two approaches to language modelling (3) are used in the hybrid system, a simple n-gram model and recurrent neural network lattice rescoring, the latter garnering impressive performance improvements. The system is evaluated using a test set that is comprised of both native and non-native speakers, which allows for some inferences to be made on the performance of the system on both cohorts.