AI/ML Investigator II — Regulatory Biomedical AI

Cognizance Technologies LLC; Silver Spring (MD); Hybride; USD 120 000 - 180 000; Plein temps

Avantages offerts par ce poste

PPO/HMO Health Plan

Résumé du poste

Cognizance Technologies, LLC in Silver Spring, MD is seeking scientists with AI/ML expertise to design, validate, and deploy AI-enabled review tools for regulatory and biomedical use. You will work on horizon scanning, data analysis workflows, and responsible AI with emphasis on auditability and traceability.

This full-time role focuses on high-stakes, compliant AI deployment. Candidates should have hands-on experience with LLMs, RAG, multi-agent orchestration, and documentation of risk, with a

Qualifications

• Experience designing and implementing LLM and agentic AI systems, including RAG, tool/function calling, multi-agent orchestration, prompt and workflow design, evidence-grounded response validation, hallucination mitigation, fallback routing, and failure analysis.
• Experience developing trustworthy, audit-focused AI workflows with audit logging, traceable workflow design, execution monitoring, source-linked evidence artifacts, run metadata, model risk documentation.
• Knowledge of model evaluation and validation practices, including data quality checks, robustness testing, drift or failure monitoring, human review controls, and documentation of system assumptions, dependencies, limitations, and risk mitigations.
• Ability to translate reviewer and scientific user needs into practical AI-assisted workflows while maintaining appropriate controls for high-stakes regulatory use.

Responsabilités

• Design, develop, test, and document AI solutions to support drug submission review workflows.
• Develop and integrate human-in-the-loop review features for AI solutions, including reviewer feedback capture, source verification, confidence or risk flagging, issue escalation, editable outputs, audit trails, and mechanisms for scientific review and approval before downstream use.
• Build pipelines to ingest, parse, structure, and analyze regulatory, scientific, and clinical documents and data sources, including PDFs, tables, figures, metadata, and structured or unstructured text, while preserving traceability to source evidence.
• Establish evaluation and validation protocols to assess AI output accuracy, reproducibility, robustness, source attribution, hallucination risk, tool-use reliability, and operational limitations; identify unexpected results and provide potential causes and mitigation strategies.
• Implement logging, monitoring, and quality assurance controls for AI workflows, including tool/function call validation, data quality checks, execution monitoring, failure reporting, token or resource usage tracking, and reproducible artifact generation.

Connaissances

AI/ML design

RAG workflows

Tool calling

Multi-agent systems

Prompt design

Evidence validation

Audit logging

Outils

Python

ML frameworks

Description du poste

Cognizance Technologies, LLC | Full time

Silver Spring, United States | Posted on 09/18/2026

To address rapidly changing program needs across several ongoing research projects within the Office of Clinical Pharmacology (OCP), the program requires the services of scientists with experience across several different scientific disciplines including bioanalysis, pharmacokinetics / pharmacodynamics (PK/PD), toxicology, immunology, microbiology, cell and molecular biology, mathematical modeling, and structure activity relationships and artificial intelligence / machine learning (AI/ML) to assist with these projects.

The objective is to provid e the program with the services of scientists with expertise in AI (Artificial Intelligence) and ML (Machine Language) to integrate AI and automation into horizon scanning and data analysis workflows, driving greater efficiency and quality in regulatory reviews and research. The scientists shall have expertise in AI and ML for assistance with validating, documenting, and deploying AI/ML, large language model (LLM), retrieval-augmented generation (RAG), and/or agentic workflow systems in biomedical, clinical, regulatory, or other high-stakes settings.

Requirements

Educational and Professional Experience

• Experience designing and implementing LLM and agentic AI systems, including RAG, tool/function calling, multi-agent orchestration, prompt and workflow design, evidence-grounded response validation, hallucination mitigation, fallback routing, and failure analysis.
  
• Experience developing trustworthy, audit-focused AI workflows with audit logging, traceable workflow design, execution monitoring, source-linked evidence artifacts, run metadata, model risk documentation, and reproducible technical documentation.
  
• Knowledge of model evaluation and validation practices, including data quality checks, robustness testing, drift or failure monitoring, human review controls, and documentation of system assumptions, dependencies, limitations, and risk mitigations.
  
• Ability to translate reviewer and scientific user needs into practical AI-assisted workflows and user-facing prototypes while maintaining appropriate controls for non-public information and high-stakes regulatory use.
  

Responsibilities

• Design, develop, test, and document AI solutions to support drug submission review workflows, including document triage, evidence extraction, literature and submission information retrieval, question answering, summarization, and generation of traceable review artifacts.
  
• Develop and integrate human-in-the-loop review features for AI solutions, including reviewer feedback capture, source verification, confidence or risk flagging, issue escalation, editable outputs, audit trails, and mechanisms for scientific review and approval before downstream use.
  
• Build pipelines to ingest, parse, structure, and analyze regulatory, scientific, and clinical documents and data sources, including PDFs, tables, figures, metadata, and structured or unstructured text, while preserving traceability to source evidence.
  
• Establish evaluation and validation protocols to assess AI output accuracy, reproducibility, robustness, source attribution, hallucination risk, tool-use reliability, and operational limitations; identify unexpected results and provide potential causes and mitigation strategies.
  
• Implement logging, monitoring, and quality assurance controls for AI workflows, including tool/function call validation, data quality checks, execution monitoring, failure reporting, token or resource usage tracking, and reproducible artifact generation.
  
• Prepare technical documentation, workflow descriptions, risk and limitation summaries, user guidance, progress reports, and other deliverables needed to support transparent development, human oversight, and future maintenance of AI-enabled review tools.
  
• PPO/HMO Health Plan (includes medical, dental, and vision)