Agentic AI Research Engineer for Planning & Support

Code Metal; San Francisco (CA); Hybrid; USD 120,000 - 160,000; Full time

Benefits offered by this job

Health care plan with 100% premium coverage

401k with 5% matching

Uncapped vacation

Flexible hybrid or remote work arrangement

Job summary

Code Metal, based in San Francisco, is looking for an Applied AI Research Engineer to design and build advanced AI systems that assist military decision-making. This role combines AI research with practical applications, focusing on creating human-machine collaboration models that ensure the reliability of decision-support systems.

The ideal candidate will have a strong tech background, preferably with a degree in Computer Science or a related field, and experience in building AI systems using modern tooling like PyTorch.

Qualifications

• 3+ years building AI, machine learning, or applied research systems.
• Experience building agentic AI systems or multi-step reasoning pipelines.
• Strong understanding of benchmarking and model evaluation.

Responsibilities

• Design and build agentic AI systems for decision support.
• Develop AI pipelines integrating models and external tools.
• Fine-tune and evaluate models for planning and decision tasks.

Skills

Python engineering skills

Experience with PyTorch

Fine-tuning language models

Multi-agent AI systems

Experiment design and evaluation

Education

Bachelor's or Master's in Computer Science or related field

Tools

Transformers

Hugging Face

Job description

Code Metal, based in San Francisco, is looking for an Applied AI Research Engineer to design and build advanced AI systems that assist military decision-making. This role combines AI research with practical applications, focusing on creating human-machine collaboration models that ensure the reliability of decision-support systems.

The ideal candidate will have a strong tech background, preferably with a degree in Computer Science or a related field, and experience in building AI systems using modern tooling like PyTorch.