Sr. Model Engineer, Post-Training & RL

SpaceXSI•Palo Alto, CA

2d•$180,000 - $600,000

About The Position

The Sr. Model Engineer, Post-Training & RL will work on SpaceXSI’s most critical post-training and reinforcement learning challenges. Areas of focus include reward modeling, preference optimization (RLHF/DPO), and reinforcement learning to improve reasoning, truthfulness, and real-world capabilities. SpaceXSI is building a Super Intelligence stack spanning supercomputers, Grok models, and applications including Grok Bot, Grok Build, Grok Imagine, Grok Voice, and X.

Requirements

Experience writing production-quality code in Python, PyTorch, JAX, Rust, or C++.
Experience training large models, such as through pre-training, post-training, fine-tuning, or reinforcement learning of neural networks with 1B+ parameters.

Nice To Haves

A strong interest in truth-seeking AI and its challenges.
Strong interest in building useful models through post-training and reinforcement learning techniques.
Experience using AI models and eagerness to advance reinforcement learning and alignment methods.
Prior experience in post-training, RLHF, or training models used by millions of people is a plus, but relevant experience is not required.
Taking pride in work and thriving in meritocratic environments.

Responsibilities

Work on critical post-training and reinforcement learning challenges.
Develop approaches involving reward modeling and preference optimization, including RLHF and DPO.
Use reinforcement learning to improve models’ reasoning, truthfulness, and real-world capabilities.