Applied AI Scientist – R&D and Regulatory

Socket.dev; Bridgewater (MA); On-site; USD 120,000 - 170,000; Full time

Job summary

Amneal is seeking an Applied AI Scientist in R&D and Regulatory to advance NLP, document intelligence and generative AI for regulatory submissions, with validated systems compliant with GxP and 21 CFR Part 11.

You will build models analyzing historical health authority correspondence, develop retrieval and generation pipelines over controlled scientific content, and support drafting and comparison for labeling.

Responsibilities

• Build models that analyze historical health authority correspondence to identify deficiency patterns and flag submission risk.
• Develop retrieval and generation pipelines over controlled scientific content with full traceability.
• Enable AI-assisted drafting and comparison for product labeling, including structured labeling content.
• Design evaluation frameworks for generative systems and automate regression testing.
• Validate AI systems under GxP AI validation framework with risk classification and intended-use definitions.
• Maintain production data pipelines on Databricks over unstructured documents with OCR and entity extraction.
• Establish benefit baselines with Regulatory and R&D owners and report on cycle-time improvements.
• Present validation evidence and model limitations to Regulatory and R&D leadership.
• Educate the organization on the usage and application of agentic systems.
• Develop literature, patent, and prior-art mining capabilities to support portfolio decisions.

Job description

The Applied AI Scientist — R&D and Regulatory builds and deploys AI systems that accelerate product development and regulatory submissions. The role applies natural language processing, document intelligence, and generative AI to Amneal's scientific and regulatory content, delivering validated systems that meet GxP and 21 CFR Part 11 requirements and are defensible under health authority inspection. The role sits within the Enterprise AI Transformation function, working in close partnership with Regulatory Affairs and R&D.

Essential Functions:

Regulatory & Scientific Agents

• Build models that analyze historical health authority correspondence — deficiency letters, information requests, and complete response letters — to identify recurring deficiency patterns and flag submission risk prior to filing.
• Develop retrieval and generation agentic systems over controlled scientific content including specifications, analytical methods, stability data, prior filings, and eCTD, with full traceability from output to source.
• Build AI-assisted drafting and comparison capability for product labeling, including structured product labeling content and reference listed drug comparison.

AI Solution Development

• Design and maintain formal evaluation frameworks for all generative systems, including curated golden datasets, defined failure taxonomies, and automated regression testing.
• Execute AI system validation under Amneal's GxP AI validation framework: risk classification, intended-use definition, qualification testing, and periodic review.
• Build and maintain production data pipelines on Databricks over unstructured scientific documents, including OCR, parsing, and entity extraction from legacy PDF sources.

Business Impact

• Establish benefit baselines with Regulatory and R&D functional owners and measure realized cycle-time and capacity outcomes.
• Present technical approaches, validation evidence, and model limitations to Regulatory, and R&D leadership.
• Enable larger organization and educate to upskill on usage and application of agentic systems.

Research and Portfolio Insights

• Develop literature, patent, and prior-art mining capabilities supporting product development screening and portfolio prioritization.