Post-Training AI Researcher — Enterprise LLM Alignment

Distyl; New York, San Francisco (NY, CA); Hybrid; USD 150,000 - 250,000; Full time

Benefits offered by this job

Equity

Medical insurance

Flexible time off

401(k) plan

Commuter benefits

In-office lunches

Access to AI models

High-impact projects

Mission-driven culture

Job summary

Distyl AI is seeking researchers to join our Post-Training team, translating foundation models into enterprise-ready systems. You will work on supervised fine-tuning, preference optimization, and continual adaptation to align models with Distyl’s workflows.

You will shape how AI-native operations are deployed across industries, building prototypes and conducting experiments that demonstrate impact using real-world data.

Qualifications

• Deep understanding of post-training techniques including supervised fine-tuning, preference optimization (RLHF/DPO), LoRA/PEFT, and instruction-tuning pipelines.
• Experience adapting frontier models to specialized domains or behaviors through data curation or continual pretraining.
• Experience building intelligent systems using models as components rather than only training models.
• Proven track record of research results via publications or public demonstrations.
• Active use of AI tools (e.g., ChatGPT, Cursor, Perplexity) to accelerate workflows.
• Strong programming and data analysis skills with ability to prototype and run experiments.

Responsibilities

• The Post-Training team adapts foundation models to real-world performance and alignment, developing techniques to align models with enterprise systems.
• Researchers investigate alignment methods, considering generalization vs. specialization and data efficiency vs. robustness.

Skills

Post-training techniques

Frontier-model adaptation

Prototyping with AI systems

Research track record

Hands-on AI tooling

Tools

Python

PyTorch

RLHF/DPO

LoRA/PEFT

Job description

Distyl AI is seeking researchers to join our Post-Training team, translating foundation models into enterprise-ready systems. You will work on supervised fine-tuning, preference optimization, and continual adaptation to align models with Distyl’s workflows.

You will shape how AI-native operations are deployed across industries, building prototypes and conducting experiments that demonstrate impact using real-world data.