oikyoni/drug-screening-llm

DrugScreenLLM — A Domain-Specific LLM for Workplace Drug Screening Compliance

Model Summary

DrugScreenLLM is a fine-tuned version of Meta's Llama 3.2 3B Instruct, specialized for workplace drug screening compliance and interpretation. It is designed to help HR professionals, occupational health nurses, Medical Review Officers (MROs), and employers accurately navigate the complex regulatory landscape of federal and state workplace drug testing.

Problem Statement

Workplace drug screening operates under a complex, frequently updated regulatory framework spanning SAMHSA mandatory guidelines, DOT 49 CFR Part 40, FDA device regulations, and 50 varying state marijuana laws. Misinterpretation of cutoff thresholds, panel compositions, or jurisdiction-specific rules can result in legal liability, wrongful termination claims, or compliance violations. No existing AI tool has been purpose-built and grounded in this regulatory domain.

Intended Use

Intended users:

• HR professionals managing workplace drug-free policies
• Occupational health nurses interpreting screening results
• Medical Review Officers (MROs)
• Employers in DOT-regulated industries
• Compliance officers

Intended use cases:

• Answering regulatory compliance questions (SAMHSA, DOT, FDA)
• Understanding cutoff thresholds by drug class and specimen type
• Navigating state-specific marijuana testing laws
• Understanding the two-step screening and confirmation process
• MRO process and split specimen questions

Out of scope:

• This model is NOT intended for clinical diagnosis
• This model is NOT a substitute for legal advice
• This model should NOT be used for individual employee medical decisions
• Results should always be verified against current Federal Register notices

Training Data

Fine-tuned on the Drug Screening Compliance Benchmark Dataset, a curated dataset of expert-crafted Q&A pairs grounded exclusively in publicly available regulatory and peer-reviewed sources including:

• SAMHSA Mandatory Guidelines for Federal Workplace Drug Testing Programs (Urine and Oral Fluid, 2023–2025)
• DOT 49 CFR Part 40 (current)
• FDA drug testing device regulations
• StatPearls peer-reviewed toxicology literature
• State marijuana workplace testing laws (all 50 states)

All training data is traceable to primary sources. No proprietary or confidential data was used.

Training Approach

• Base model: meta-llama/Llama-3.2-3B-Instruct
• Method: QLoRA (Quantized Low-Rank Adaptation)
• Framework: Hugging Face PEFT + Transformers
• Hardware: Kaggle T4 GPU
• Train/validation split: 90% / 10%

Evaluation Results

Metric	Value
Training Loss	0.478312 (Epoch 3)
Validation Loss	0.682971 (Epoch 2)
Best Mean Token Accuracy	83.0%

Limitations

• Regulatory cutoffs and panel compositions change frequently. Always verify against the current Federal Register notice.
• State marijuana laws are evolving rapidly. Consult current state statutes before making employment decisions.
• This model reflects training data as of June 2026.
• The model may not reflect regulatory changes published after the training data cutoff.

Regulatory Notice

Drug screening decisions with legal or employment consequences must be reviewed by a qualified Medical Review Officer (MRO) and legal counsel. This model is an informational tool only.

Citation

If you use this model or the associated benchmark dataset in your research, please cite:

@misc{drugscreenllm2026, title={ DrugScreenLLM: A Domain-Specific Language Model for Workplace Drug Screening Compliance }, author={ Sairth Bhattacharjya }, year={ 2026 }, publisher={ HuggingFace }, url={ https://huggingface.co/oikyoni/drug-screening-llm } }

Author

Developed by Sairath Bhattacharjya, a workplace drug screening domain expert with professional experience in federally regulated drug testing programs.