Senior AI Engineer, Post-Training for Legal AI

Carta; San Francisco, New York (CA, NY); Hybrid; USD 242,000 - 285,000; Full time

Benefits offered by this job

Equity

Excellent benefits

Commission plans

Job summary

Carta is hiring an AI Engineer to lead model-centric projects, from data to production. You will post-train models, design training objectives, and build the systems that bring AI capabilities to users. You will collaborate with engineers and legal teams to translate workflows into model and tool decisions.

The role emphasizes cross-functional ownership, end-to-end training, and deployment within a hybrid work model in San Francisco and other locations.

Qualifications

• Hands-on experience with post-training of large language models using PyTorch or similar frameworks.
• Ability to build products around models, including agents, tools, services, and production infrastructure.
• Stays current on open-weight models and post-training techniques.
• Experience owning model development or post-training work and shipping AI systems.

Responsibilities

• Lead technically complex, model-centric projects across the training and evaluation lifecycle.
• Build and optimize training datasets, labeling guidance, and human feedback loops.
• Own the training stack and diagnose distributed training performance.
• Develop production-ready systems for model serving, agents, and evaluation pipelines.
• Collaborate with product and engineering teams on model and workflow co-design.
• Translate real workflows from lawyers and domain experts into model decisions.

Skills

LLM post-training

PyTorch

Model deployment

Data pipelines

Distributed training

Tools

PyTorch

Job description

Carta is hiring an AI Engineer to lead model-centric projects, from data to production. You will post-train models, design training objectives, and build the systems that bring AI capabilities to users. You will collaborate with engineers and legal teams to translate workflows into model and tool decisions.

The role emphasizes cross-functional ownership, end-to-end training, and deployment within a hybrid work model in San Francisco and other locations.