Distyl AI
Applied AI Researcher, Post-Training
$150K–$250K
San Francisco +1 · Hybrid · Full-time · Posted Oct 16, 2025
- Location
- San Francisco
New York
About the job
Researches and develops post-training methods and AI systems that adapt foundation models to enterprise performance, alignment, and reliability requirements.
What you'll do
- Develop and evaluate supervised fine-tuning, preference optimization, and continual adaptation techniques.
- Investigate methods to align large models with human and system-level objectives, balancing generalization, robustness, and controllability.
- Build and test prototypes and experiments to demonstrate the effectiveness of AI methods.
What they ask for
- Familiarity with supervised fine-tuning, preference optimization (RLHF/DPO), LoRA/PEFT, and instruction-tuning pipelines.
- Experience tuning or adapting LLMs/SLMs for specialized domains or behaviors, using data curation, reward modeling, or continual pretraining.
- Research results demonstrated through published or other work.
- Strong programming and data analysis skills to build prototypes and run experiments.
Nice to have
- Expertise in compound AI systems, agentic collaboration, and techniques such as ensembling, ReAct, and graph-of-thoughts.
- Daily use of AI tools such as ChatGPT, Cursor, and Perplexity.
Pay
$150K – $250K (base salary)
Tools
RLHF, DPO, LoRA/PEFT, LLMs/SLMs, ChatGPT, Cursor, Perplexity, ReAct
AI focus
Post-training and adapting foundation models for enterprise use, including supervised fine-tuning, preference optimization, continual adaptation, alignment research, and compound AI systems.
Model-written summary · the original posting is authoritative