To strengthen our large model capabilities, Stellar AGI Labs spent the first half of 2026 learning the latest model technologies around the world. Following organizational adjustments in the first half of the year, we officially launched the M&N Large Model R&D Program on May 1, 2026.
Program Overview
The M&N program aims to build three core models:
**General-Purpose Large Model** — a practical foundation model for broad understanding and generation**Multimodal Large Model** — fusing text, image, and voice for cross-modal intelligence**Hong Kong Language Large Model** — dedicated to Hong Kong's indigenous language, with authentic local linguistic understandingWe are starting in Hong Kong and Singapore, and will gradually invest more resources as the program scales.
Open Positions
LLM Trainer / Large Model Training Specialist
Design and execute training pipelines for general-purpose, multimodal, and Hong Kong language modelsBuild and curate high-quality datasets across text, image, audio, and local dialect corporaEvaluate model outputs, design benchmarks, and drive continuous quality improvementCollaborate with research engineers to optimize training efficiency and model performanceRelated Positions
**Data Annotation Lead** — manage annotation teams and quality standards for SFT and local language data**ML Research Engineer** — distributed training, inference optimization, and model architecture research**Multimodal Engineer** — vision-language fusion, speech interaction, and cross-modal alignment**Hong Kong Language Data Specialist** — curate and validate Hong Kong colloquial and indigenous language training data**Prompt Engineer** — prompt design, evaluation frameworks, and application integrationRequirements
Hands-on experience with LLM training, fine-tuning, or data engineering (strong AI fundamentals and fast learners welcome)Familiarity with PyTorch, Transformers, DeepSpeed, or similar frameworksStrong sense of data quality and model behavior; detail-orientedBased in Hong Kong or Singapore, or willing to relocateWhat We Offer
**Compensation: negotiable** — competitive packages for the right talentDirect involvement in building three large models from the ground upDual-hub working environment across Hong Kong and SingaporeFlexible arrangements and a fast-moving, engineering-driven cultureHow to Apply
Send your resume (and anything that demonstrates your work — papers, projects, GitHub) to [email protected] with the subject line: `Application - [Position] - [Your Name]`.
We review every application and respond to qualified candidates within 5 business days.