Senior Model Engineer Mid-Training
SpaceXAI builds frontier AI models and products including the Grok conversational assistant and a multimodal developer API.
Funding history
About SpaceXAI
SpaceXAI is a US-based AI company focused on accelerating scientific discovery and understanding the universe through reasoning, voice, image, video, and generative AI systems. Formerly xAI, it was acquired by SpaceX effective February 2, 2026, while continuing to operate the x.ai site and services.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will scale synthetic coding data, distill flagship-model intelligence into flash models, optimize mid-training data mixtures, create long-context data recipes, and develop robust evaluations for mid-training checkpoints.
Requirements
- Production-quality programming in Python, PyTorch, JAX, Rust, or C++
- Experience training large models, including pre-training, post-training, fine-tuning, or reinforcement learning training of neural networks with more than 1 billion parameters
- Scaling laws
- Machine learning experimentation
- AI training data curation
- Large-scale data processing
- Machine learning scaling
- Spark
- Ray
Responsibilities
- Scale synthetic coding data to trillions of tokens using large-scale Docker verification
- Distill flagship-model intelligence into flash models through synthetic data generation
- Optimize mid-training data mixtures to raise the ceiling for reinforcement learning
- Engineer long-context data recipes
- Develop robust and diverse evaluations for mid-training checkpoints
Benefits
- Equity
- Comprehensive medical coverage
- Vision coverage
- Dental coverage
- 401(k) retirement plan
- Short-term disability insurance
- Long-term disability insurance
- Life insurance
- Discounts and perks