AI Model Trainer

Gateway Search; Singapore; On-site; SGD 70,000 - 100,000; Full time

Job summary

Gateway Search is seeking an AI Model Trainer in Singapore to support the development, training, evaluation, and optimization of small, domain-specific AI models. You will curate datasets, annotate data, and collaborate with engineering and product teams to translate business needs into training data and evaluation metrics.

The role emphasizes data quality, model evaluation, and continuous optimization across NLP and ML tasks, with opportunities to influence model performance in real-world

Qualifications

• Bachelor's degree or above in Computer Science, Artificial Intelligence, Data Science, Mathematics, Statistics, Engineering, Linguistics, Cognitive Science, or related discipline.
• Approximately 1–3+ years of relevant experience in AI model training, machine learning data preparation, data annotation, model evaluation, algorithm testing, NLP, CV, or related field.
• Good understanding of fundamental ML concepts including training/validation/test datasets, model accuracy, precision, recall, and overfitting.
• Hands-on experience preparing, labeling, reviewing, cleaning, or improving datasets for AI/ML model training.

Responsibilities

• Develop, curate, clean, and refine high-quality datasets for small, domain-specific AI models.
• Support model training, fine-tuning, testing, and optimization for business applications.
• Annotate, review, and validate training data for accuracy and consistency.
• Evaluate model outputs against quality standards: accuracy, relevance, consistency, robustness, task completion.
• Design evaluation criteria, test datasets, benchmarking methodologies, edge cases and failure scenarios.
• Identify recurring model issues and provide structured recommendations for improvement.
• Analyze relationship between training data quality and model performance to refine datasets.
• Develop and maintain annotation guidelines, labeling standards, and QC processes.
• Support workflows such as supervised fine-tuning, incremental training, data augmentation, and synthetic data generation.
• Collaborate with AI/ML engineers and product teams to translate requirements into datasets and evaluation metrics.
• Conduct continuous model testing across normal and edge cases to improve reliability.
• Track model performance across iterations and report improvements and regressions.
• Stay up to date with developments in small language models, domain-specific models, ML evaluation methodologies.

Skills

Data annotation

NLP

Model evaluation

Python scripting

Critical thinking

Communication

Education

Bachelor's in CS/AI/Data Science

Tools

PyTorch

TensorFlow

Hugging Face

scikit-learn

SQL

Job description

Our client is a fast-growing AI technology company developing advanced, human-centric artificial intelligence solutions. The team combines machine learning, natural language processing, and personalized analytical frameworks to build AI systems that can understand complex user contexts, behaviors, and decision-making patterns.

As the business continues to expand its AI capabilities, they are looking for an AI Model Trainer to support the development, training, evaluation, and continuous optimization of small, domain-specific, and task-oriented AI models.

This role is ideal for someone who enjoys working at the intersection of AI model training, data quality, model evaluation, and applied machine learning, and who is comfortable collaborating closely with engineering and product teams in a fast-paced environment.

Key Responsibilities

• Develop, curate, clean, and refine high-quality datasets for small, domain-specific, and task-oriented AI models.
• Support model training, fine-tuning, testing, and optimization for specific business applications, including classification, information extraction, recommendation, prediction, recognition, and conversational AI.
• Annotate, review, and validate training data to ensure accuracy, consistency, completeness, and alignment with model requirements.
• Evaluate model outputs against defined quality standards, including accuracy, relevance, consistency, robustness, and task-completion quality.
• Design evaluation criteria, test datasets, benchmarking methodologies, edge cases, and failure scenarios to assess model performance.
• Identify recurring model issues, including misclassifications, false positives/negatives, data bias, performance gaps, and failure patterns, and provide structured recommendations for improvement.
• Analyze the relationship between training data quality and model performance, and continuously refine datasets based on evaluation findings.
• Develop and maintain annotation guidelines, labeling standards, training instructions, and quality-control processes.
• Support workflows such as supervised fine-tuning, incremental training, data augmentation, synthetic data generation, and other model optimization approaches.
• Work closely with AI/ML engineers, algorithm engineers, and product teams to translate business requirements into training datasets, labeling rules, evaluation metrics, and model improvement strategies.
• Conduct continuous model testing across normal scenarios, edge cases, and real-world business use cases to improve model reliability and generalization.
• Track model performance across training iterations and provide clear analysis of improvements, regressions, and recommended next steps.
• Stay up to date with developments in small language models (SLMs), domain-specific AI models, machine learning, model fine-tuning, data annotation, and AI evaluation methodologies.

Requirements

• Bachelor's degree or above in Computer Science, Artificial Intelligence, Data Science, Mathematics, Statistics, Engineering, Linguistics, Cognitive Science, or a related discipline.
• Approximately 1–3+ years of relevant experience in AI model training, machine learning data preparation, data annotation, model evaluation, algorithm testing, NLP, computer vision, or a related field.
• Good understanding of fundamental machine learning concepts, including training/validation/test datasets, model accuracy, precision, recall, and overfitting.
• Hands‑on experience preparing, labeling, reviewing, cleaning, or improving datasets used for AI or machine learning model training.
• Experience evaluating model predictions or outputs and diagnosing model performance issues is highly preferred.
• Strong analytical and critical‑thinking skills with excellent attention to detail.
• Ability to identify subtle differences in model outputs, recognize recurring error patterns, and clearly explain potential causes of model failures.
• Ability to develop clear labeling guidelines, evaluation standards, and quality‑control frameworks.
• Strong written and verbal communication skills.
• Comfortable collaborating closely with technical teams and iterating quickly based on model performance and business feedback.
• Able to thrive in a fast‑paced environment where product requirements and model priorities may evolve rapidly.

Preferred Qualifications

• Experience working with small language models (SLMs), domain‑specific models, lightweight AI models, or task‑specific machine learning models.
• Hands‑on exposure to model fine‑tuning, supervised learning, transfer learning, parameter‑efficient fine‑tuning (PEFT), or incremental model training.
• Experience using Python, SQL, or other programming/scripting languages for data processing, analysis, automation, or model evaluation.
• Familiarity with machine learning frameworks and tools such as PyTorch, TensorFlow, Hugging Face, scikit‑learn, or similar technologies.
• Experience with data‑labeling platforms, dataset‑management tools, or model‑evaluation frameworks.
• Experience building or maintaining training datasets, validation datasets, test sets, annotation guidelines, or benchmarking systems.
• Familiarity with techniques such as data augmentation, synthetic data generation, hard‑negative mining, active learning, or error‑driven dataset iteration.
• Experience working with NLP, computer vision, speech recognition, recommendation systems, classification models, or other domain‑specific AI applications.
• Academic research, personal projects, or practical project experience in machine learning, NLP, computer vision, data science, or related AI fields would be an advantage.

Why Consider This Opportunity?

You will have the opportunity to work directly on practical AI-model development and improvement, with significant exposure to the full model iteration cycle — from data preparation and training through evaluation, error analysis, and continuous optimization.

The role offers strong learning potential for candidates looking to build deeper expertise in applied AI, small/domain‑specific models, model evaluation, and data‑centric AI development.

All applications will be treated with strictest confidence. We regret to inform you that only shortlisted candidates will be notified.

EA Licence No: 19C9807

Registration No: R22110556 (ZHOU DONGYANG)