MTS, Post-Training (Enterprise)

What you'll do

• Fine-tune and align models for enterprise domains and production requirements.
• Build evaluation suites, curate datasets, and monitor for regressions.
• Work with customers and stakeholders to translate requirements into metrics and explain model behavior.

What they ask for

• Have post-trained at least one LLM deployed to production users and measured its success.
• Have end-to-end experience building benchmarks and evaluation datasets and agreeing on metrics with stakeholders.
• Understand regression risks and how to preserve existing model capabilities during training.
• Have strong software and ML fundamentals, including production data pipelines and post-training frameworks.

Nice to have

• Experience with conversational or task-oriented assistants, LLM judges or reward models, or production data flywheels.
• Hands-on experience with open-weight models or a background as an FDE, founder, or early-stage startup employee.

Pay

$300,000 – $350,000 USD / year (base salary); 25% performance-based bonus + equity (additional compensation; amount not stated)

Tools

Llama, Qwen, Mistral, DeepSeek

AI focus

Enterprise post-training, fine-tuning, alignment, evaluation, and deployment of models; includes preference tuning, data curation, and regression mitigation.