Senior AI Engineer Post Training
Carta provides an intelligent platform for private capital operations, connecting data, workflows, and people.
About Carta
Carta operates a platform for private capital that helps companies, private equity and venture capital firms, limited partners, and other financial organizations manage portfolios, move money, and model investment scenarios. Its offerings include equity and cap table management, valuations, compensation, liquidity, fund administration, fund tax, SPVs, deal CRM, portfolio valuations, loan operations, LP analytics, compliance, contracts, and attorney-led legal services. Carta combines software, AI capabilities, and expert human services for private capital workflows.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will lead model-centric projects from post-training through production. You will develop and evaluate language models, build training data and pipelines, operate training and serving infrastructure, and partner with engineers and legal experts to turn workflows into model and system decisions.
Requirements
- Hands-on experience with LLM post-training using PyTorch or equivalent frameworks
- Knowledge of training, evaluation, and inference systems
- Experience building agents, tools, services, and production infrastructure
- Experience owning model development or post-training work in applied settings
- Experience shipping AI systems to real users
- Strong judgment in model selection, data, training objectives, and evaluation
- Experience spanning model-level training and production engineering systems
Responsibilities
- Post-train open-weight language models on proprietary legal data
- Design training objectives, select base models, train, evaluate, and iterate on models
- Apply supervised fine-tuning, preference optimization, reinforcement learning, and related methods
- Build training datasets, data pipelines, labeling guidance, model-generated data, and human-feedback loops
- Own and optimize the training stack and distributed training runs
- Build and operate model serving, agents, evaluation pipelines, and supporting infrastructure
- Partner on model and system co-design
- Translate legal workflows into model, data, and evaluation decisions
Benefits
- Equity for full-time roles
