Senior LLM Research Scientist | Remote

Requirements

• Train transformer-based language models from scratch and fine-tune open-weight models across data- and compute-constrained regimes, including fixed-budget allocation across model size, tokens, and epochs
• Construct and curate pre-training corpora from raw web crawls including filtering, deduplication, quality classification, and mixture optimization with rigorous measurement of data interventions
• Build post-training pipelines using supervised fine-tuning, synthetic dataset construction, rejection sampling, or preference optimization methods (DPO, RLHF, RLAIF)
• Diagnose and resolve training instabilities, optimization failures, and convergence issues in large-scale model runs
• Apply alignment fine-tuning techniques to shape refusal behavior, truthfulness, and reasoning quality without degrading general capability
• Demonstrated research track record via a degree from a top-100 university, experience at a leading AI lab or major tech company, or impactful publications or open-source contributions
• Proficiency in PyTorch, JAX, TensorFlow, or comparable ML frameworks; experience with scaling laws, curriculum learning, LLM evaluation, or RL for language models is a plus

Qualifications

Competitive hourly rate of $100 $120, paid weekly via Stripe or Wise as an independent contractor

Full flexibility in schedule, work on your own time, with projects that can scale up or down based on performance and scope

Engaged as an independent contractor; no access to confidential third-party information required

Ongoing pipeline of research projects across pre-training, data curation, and post-training domains