Senior Staff Engineer, Multimodal AI Pretraining

Hark, Inc.; San Jose (CA); On-site; USD 180,000 - 450,000; Full time

Job summary

Hark, Inc. is seeking an AI Developer to join their Omni team, focusing on multimodal intelligence across text, audio, and vision. Candidates will develop large-scale pretraining systems and work on improving foundational models.

The ideal candidate will demonstrate expertise in distributed training, data pipelines, and optimization techniques. Competitive compensation ranges from $180,000 to $450,000 annually based on experience, and additional benefits may apply.

Qualifications

• Proven track record of improving large-scale neural network performance.
• Deep understanding of LLM or multimodal pretraining.
• Experience in data-driven experimentation and debugging at scale.

Responsibilities

• Drive research and development in large-scale LLM and multimodal pretraining.
• Develop and optimize data pipelines for pretraining.
• Design and implement efficient training strategies for foundation models.
• Build and improve pretraining infrastructure.
• Develop evaluation frameworks to measure pretraining progress.
• Collaborate with teams to push the frontier of model performance.

Skills

Large-scale distributed training

Neural network optimization

Data-driven experimentation

Scaling behavior understanding

Ownership mindset

Education

Experience with high-performance computing

Experience with multimodal pretraining

Job description

Hark, Inc. is seeking an AI Developer to join their Omni team, focusing on multimodal intelligence across text, audio, and vision. Candidates will develop large-scale pretraining systems and work on improving foundational models.

The ideal candidate will demonstrate expertise in distributed training, data pipelines, and optimization techniques. Competitive compensation ranges from $180,000 to $450,000 annually based on experience, and additional benefits may apply.