LLM Fine-Tuning & Alignment Specialist

About the opportunity

Advance Constitutional AI, RLHF, and automated prompt red-teaming across Claude model releases.

What you will do

• check_circle Curate high-quality instruction datasets and formulate constitutional alignment critique principles.
• check_circle Train and evaluate Direct Preference Optimization (DPO) and PPO alignment models at scale.
• check_circle Build automated red-teaming agents that probe safety boundaries, prompt injections, and jailbreaks.
• check_circle Collaborate with interpretability researchers to inspect internal neural activations during reasoning.

What we are looking for

• arrow_circle_right Strong track record in deep learning with specific focus on Large Language Model fine-tuning and alignment.
• arrow_circle_right Experience with PyTorch, DeepSpeed, Megatron, DPO, and distributed training clusters.
• arrow_circle_right Passionate dedication to AI safety, truthfulness, and ethical technology development.

Skills & Tech Stack