AI Engineer

This is a fully remote position, open to applicants in Tennessee.

📋 Description

• Collaborate with domain experts to capture specialized knowledge in a structured and usable format.

• Evaluate source materials and datasets for quality, consistency, and coverage.

• Conduct exploratory and statistical analyses to guide data, modeling, and evaluation decisions.

• Generate and maintain training and evaluation datasets through both automated and human validation processes.

• Develop reinforcement learning environments, reward functions, and verifiers.

• Train and fine-tune language models utilizing supervised and reinforcement learning techniques, including PPO, DPO, and GRPO.

• Set baselines, execute ablation studies, and troubleshoot reward hacking, instability, and data leakage issues.

• Stay updated with research literature and implement promising methodologies.

• Oversee versioned configurations, reproducible runs, experiment tracking, and compute budgeting.

• Create and enhance evaluation frameworks, metrics, and benchmarks.

• Evaluate models and methodologies through structured experiments and A/B testing.

• Document outcomes, setbacks, and subsequent actions in technical reports.

• Share findings with engineers, domain experts, and business leaders.

• Take ownership of experiments from design through coding, training, and analysis of results.

• Work closely with domain experts and adjust data, training, and evaluation strategies as knowledge progresses.

• Ensure compliance with HIPAA and EnableComp data policies.

⛳️ Requirements

• Bachelor’s degree in computer science, machine learning, statistics, or a related discipline; MS or PhD is preferred.

• Equivalent hands-on experience will also be accepted.

• Over 3 years of hands-on experience in training or fine-tuning language models.

• Experience with reinforcement learning, including end-to-end design of environments and reward functions, as well as policy training.

• Proficient understanding of PPO, DPO, or GRPO and their potential failure modes.

• Strong skills in Python and PyTorch.

• Familiarity with the Hugging Face ecosystem and efficient parameter fine-tuning techniques.

• Background in building and assessing LLM-based agents or multi-turn systems.

• Experience in designing language model evaluations that include held-out sets and automated grading.

• Capability to analyze intricate documentation and convert it into precise, structured, and testable formats.

• Proficient technical writing skills.

• Ability to work comfortably with non-machine-learning domain experts.

• Regular and punctual attendance.

• Equivalent combinations of education and experience will be taken into account.

• Adherence to HIPAA and EnableComp data policies.

• Experience with healthcare data or other highly regulated sectors is highly preferred.

• Preference for experience with structured or schema-constrained model outputs.

• Broad understanding of reinforcement learning algorithms and advanced reward design is highly preferred.

• Experience with knowledge graphs or ontologies is highly preferred.

• Preferred experience in synthetic data generation and validation.

• Publications or contributions to open-source projects in relevant fields are highly preferred.

• Experience managing PHI in compliance with HIPAA is highly preferred.

🏝️ Benefits

• Full-time employment.

• Opportunities for professional growth and development.

• Access to tools, resources, and support for career advancement.

• Commitment to an equal opportunity and discrimination-free workplace.

• Reasonable accommodations for qualified individuals with disabilities.

• Flexible remote work arrangements.