LLM Architect

BRIGHTOIL PETROLEUM (S'PORE) PTE. LTD.; Singapore; On-site; SGD 60,000 - 80,000; Full time

Job summary

An innovative firm is seeking a talented individual to design and optimize large language models. In this role, you will explore cutting-edge deep learning architectures and develop systems for efficient model training and deployment. You will also establish performance evaluation systems to monitor training metrics, ensuring high model quality. This position offers a unique opportunity to work with advanced techniques like fine-tuning and prompt engineering, contributing to the evolution of AI technologies. If you have a passion for artificial intelligence and a drive to innovate, this role is perfect for you!

Qualifications

• Bachelor's degree in relevant fields required.
• At least 1 year of experience in model training and deployment.

Responsibilities

• Design and optimize large language models and streamline training processes.
• Develop systems for efficient model training and performance evaluation.

Skills

Deep Learning

Fine-tuning Strategies

Prompt Engineering

Model Training

Resource Management

Education

Bachelor degree in Computer Science

Bachelor degree in Artificial Intelligence

Bachelor degree in Mathematics

Tools

Seq2Seq

Transformer

Job description

Descriptions:

• Design and optimize large language models, devise fine-tuning strategies, and streamline the training process.
• Explore deep learning architectures like Seq2Seq, Transformer, and advanced techniques including Fine-tuning, Prompt Engineering, and Soft Prompting (SFT).
• Develop systems for efficient model training and deployment, involving data preprocessing, parallel training, and resource management.
• Establish performance evaluation systems and monitor training metrics to ensure model quality and iteration efficiency.

Requirements:

• Bachelor degree or higher in Computer Science, Artificial Intelligence, Mathematics, or related fields.
• LLM. AIGC required.
• At least 1 year of relevant experience.