Principal Applied AI Scientist – LLM Post-Training & Evaluation

Opportunity Overview

Location: 100% Remote (United States)
Employment Type: Full-Time Direct-Hire via Accel DNA
Compensation Range: $215,000 – $255,000 Base + Significant Equity Grant
Hiring Agency: Accel DNA – Premier Frontier Data & AI Talent Partner

An elite cybersecurity defense enterprise is hiring a Principal Applied AI Scientist to head private model post-training, specialized domain adaptation, and adversarial safety alignment. You will take raw open-weight foundation models and build proprietary, industry-leading specialized models.

Key Responsibilities:

• Lead the post-training strategy for specialized open-weight models using Supervised Fine-Tuning (SFT), DPO (Direct Preference Optimization), and RLHF.
• Design programmatic synthetic data generation pipelines, quality filtering, and deduplication workflows using frontier models and rejection sampling.
• Develop comprehensive automated evaluation benchmarks measuring reasoning fidelity, hallucination rate, and adversarial safety vulnerabilities.
• Publish research findings and represent the organization at premier AI conferences (NeurIPS, ICML, ICLR).

What You Bring:

• PhD or MS in Computer Science, Machine Learning, Computational Linguistics, or related quantitative field.
• 5+ years of applied research experience with a track record of fine-tuning, aligning, and deploying large-scale neural architectures (Transformers).
• Proficiency with distributed training frameworks (DeepSpeed, Megatron-LM, FSDP, PyTorch, Hugging Face TRL).
• Deep passion for empirical evaluation and model alignment methodologies.

How to Apply:

Submit your CV and published research links to contact@accel-dna.com or connect with our executive search team at https://accel-dna.com/contact-us/.