AI/NLP Engineer – LLM Prototype for Ethics + Research Compliance (SBIR Phase I)

AI/NLP Engineer for NIH SBIR Phase I Project About the Project I am the founder of an early-stage startup developing an AI-powered platform intended to improve regulatory decision-making and document quality in clinical and human-subjects research. The project was submitted to NIH as a Phase I SBIR application. We have received an Intent to Fund notification and are currently completing NIH’s Just-in-Time requirements before a final award decision. Funding is not official until NIH issues a formal Notice of Award. Immediate Need I am seeking an experienced AI/NLP Engineer who would be interested in joining the project as Senior/Key Personnel and performing the funded development role if NIH issues the award. Because of the current NIH timeline, the selected candidate must be available this week to complete several administrative requirements: * NIH Biographical Sketch Common Form * NIH Current and Pending (Other) Support Common Form * Required Research Security Training * Basic information needed for the proposed personnel update I will provide clear instructions and support throughout the process. No technical development work will be requested before an award is issued and a paid agreement is in place. Funded Role If NIH accepts the personnel update and issues the award, the position would include: Budget: $38,350 Level of effort: 3.4 calendar months or ≈ 589 hours Project period: Approximately 6–12 months Work arrangement: Remote Most development work is expected to occur during the first several months, followed by testing, refinement, documentation, and integration support. The final milestones, schedule, payment structure, deliverables, and contract terms would be confirmed before paid work begins. Anticipated Responsibilities The AI/NLP Engineer would work with the Director of Software Engineering, Full Stack Developer, Data Architect, UX/UI Designer, and regulatory experts to: * Design and implement the Phase I AI/NLP architecture * Develop a retrieval-augmented generation system grounded in authoritative regulatory materials * Build document-ingestion, indexing, retrieval, and analysis pipelines * Analyze research protocols, informed-consent forms, and related documents * Develop structured, source-grounded outputs with traceability * Implement safeguards for hallucinations, unsupported conclusions, and prompt manipulation * Develop methods for evaluating retrieval quality, factual accuracy, robustness, and expert agreement * Integrate the AI/NLP components with the broader application * Document the architecture, evaluation methods, limitations, and results Relevant Experience Strong candidates will have experience in several of the following: * Python-based AI or NLP development * Large language models and RAG systems * Embeddings, vector databases, hybrid search, or reranking * Long-document processing and information extraction * Prompt design and structured model outputs * LLM evaluation and hallucination testing * Prompt-injection defenses and responsible AI safeguards * API development and integration * Clear technical documentation and team collaboration Experience in healthcare, clinical research, legal technology, compliance, or another accuracy-sensitive field is helpful but not required. Please briefly let me know: 1. A summary of your relevant AI/NLP, LLM, or RAG experience. 2. Whether you would be available to complete the required NIH administrative forms this week. 3. Whether you would be available to perform approximately 3.4 calendar months of effort (about 590 hours) over the Phase I project if funded. Selection and paid work are contingent on NIH accepting the personnel update and issuing a formal Notice of Award.