AI Model Trainer

About the Opportunity

Join a rapidly growing AI technology firm focused on crafting advanced, human-centered artificial intelligence solutions. This team integrates machine learning, natural language processing, and custom analytical frameworks to create AI systems capable of understanding complex user behaviors, contexts, and decision-making patterns. As this company's AI products expand, they’re seeking an AI Model Trainer to contribute to development, training, evaluation, and ongoing enhancement of specialized, small-scale domain-specific AI models.

Key Responsibilities

• Developing, curating, and refining high-quality datasets tailored for small domain-specific AI applications.
• Assisting in training, fine-tuning, testing, and optimizing models for tasks such as classification, information extraction, recommendations, predictions, recognition, and conversational AI.
• Annotating, reviewing, and validating training data to maintain accuracy, consistency, and completeness aligned with model needs.
• Assessing model outputs versus quality benchmarks including accuracy, relevance, consistency, robustness, and task completion.
• Designing evaluation frameworks with criteria, test data, benchmarking methods, edge cases, and failure scenarios to gauge performance.
• Identifying recurring model challenges like misclassifications, false positives/negatives, biases, and performance gaps, proposing structured solutions for improvement.
• Analyzing how training data quality influences model effectiveness and iteratively improving datasets accordingly.
• Creating and upholding annotation standards, labeling guidelines, training directives, and quality control processes.
• Supporting workflows involving supervised fine-tuning, incremental training, data augmentation, synthetic data, and other optimization techniques.
• Collaborating closely with AI/ML engineers, algorithm developers, and product teams to convert business goals into dataset creation, labeling, evaluation metrics, and enhancement strategies.
• Conducting ongoing model testing across standard, edge, and real-world use cases to boost model reliability and generalization.
• Monitoring model performance through training cycles and delivering clear assessments of progress and next steps.
• Keeping abreast of advances in small language models, domain-specific AI, machine learning, data annotation, and AI evaluation methods.

Required Qualifications and Experience

• Bachelor’s degree or higher in Computer Science, AI, Data Science, Mathematics, Statistics, Engineering, Linguistics, Cognitive Science, or related fields.
• Between one to three-plus years’ experience in AI model training, data preparation, annotation, model evaluation, algorithm testing, NLP, computer vision, or related areas.
• Sound understanding of machine learning concepts including dataset management, accuracy, precision, recall, and overfitting.
• Practical experience in preparing, labeling, cleaning, or improving datasets for AI/ML model training.
• Preferred experience in analyzing model predictions and troubleshooting performance.
• Excellent analytical and critical thinking with strong attention to detail.
• Ability to detect nuanced differences in outputs, identify repeated errors, and explain model failure causes clearly.
• Skill in creating labeling protocols, evaluation standards, and quality assurance frameworks.
• Effective communication abilities both written and verbal.
• Comfort collaborating with technical teams and adapting quickly in a dynamic environment with evolving requirements.

Preferred Qualifications

• Experience with small language models, domain-focus models, lightweight or task-specific AI models.
• Hands-on exposure to fine-tuning, supervised learning, transfer learning, parameter-efficient fine-tuning, or incremental training methods.
• Programming knowledge in Python, SQL, or other scripting languages for analysis, automation, or evaluation.
• Familiarity with ML frameworks like PyTorch, TensorFlow, Hugging Face, or scikit-learn.
• Experience with data-labeling tools, dataset management, or model evaluation systems.
• Track record building or maintaining training and benchmarking datasets and guidelines.
• Knowledge of data augmentation, synthetic data creation, hard-negative mining, active learning, or iterative dataset improvement techniques.
• Background in NLP, computer vision, speech recognition, recommendation systems, classification models, or other domain AI applications.
• Academic or practical projects in ML, NLP, computer vision, data science, or related AI fields is beneficial.

Why Join?

This position provides direct involvement in the full AI model development lifecycle, from data preparation and training to evaluation, error analysis, and continuous refinement. It offers valuable learning for developing deeper expertise in applied AI, small-scale models, evaluation methods, and data-centric AI enhancement.

All candidate information will be handled confidentially. Only those shortlisted will be contacted.

EA Licence No: 19C9807
Registration No: R22110556 (ZHOU DONGYANG)