Ali ShahmohammadiPh.D.

Associate Director, Applied AI Engineering & Scientific Data

See writing →

Summary

Applied AI engineer with a decade in pharmaceutical R&D, building the MCP servers, agentic pipelines, and evaluation harnesses that make scientific data trustworthy for AI. Ships production systems in regulated life-sciences environments, grounded in a background in physics-informed and mechanistic modeling.

Experience

Associate Director, Applied AI Engineering & Scientific Data

Sep 2022 – Present

Takeda Pharmaceutical Inc. · Greater Boston, MA

  • Designed and deployed a production agentic AI system for ontology curation — a multi-agent LangGraph pipeline with human-in-the-loop governance — cutting curation cycle times significantly.
  • Built MCP-style scientific data connectors integrating ELN, LIMS, and AWS into a validated, agent-ready data layer across R&D domains.
  • Architected a next-generation R&D data catalog with automated quality checks, lineage, and metadata management that gates data before it reaches AI pipelines.
  • Led enterprise FAIR data strategy and delivered FAIR Studio (Pistoia Alliance), an industry-alliance product operationalizing FAIR assessment at scale.
  • Built a digital twin for continuous manufacturing of a small-molecule API, integrating unit-operation mechanistic models.
  • Established strategic research partnerships with MIT, BYU, Brown, and Purdue on PINNs and in-silico development.

Senior CMC Scientist

Aug 2021 – Sep 2022

Moderna Inc. · Greater Boston, MA

  • Developed ML models for mRNA drug-substance and drug-product stability and shelf-life prediction (ICH-compliant).
  • Optimized the IVT reaction for mRNA process characterization using combined ML and fundamental modeling.
  • Led comparability and product-specification projects through process scale-up phases.
  • Contributed to IND and BLA submissions through statistical analysis and documentation.

Post-Doctoral Research Fellow

Oct 2019 – Aug 2021

The University of Texas at Austin · Austin, TX

  • Developed fundamental models for thin-film gallium phosphate on silicon for plasma-etch optimization.
  • Built model-based DoE tools in R Shiny for plasma-etch process optimization.
  • Created a first-principles model for the viscoelastic properties of adhesive soft particles.
  • Performed molecular-dynamics simulations with 100,000+ particles.

Education

Ph.D., Chemical Engineering

Queen's University · Kingston, ON, Canada

Process systems engineering & statistical design of experiments — developed a new design-of-experiments algorithm for building mathematical models via a process-automation system.

2019  (verify year)

M.Sc., Chemical Engineering

Tarbiat Modares University · Tehran, Iran

2013

B.Sc., Chemical Engineering

Tehran University · Tehran, Iran

2011

Selected Writing

Agentic AI for Data Governance: A Production Pattern for First-Pass Curation
2026 · agentic systems, PROV-O provenance, human-in-the-loop
From Ontology to MDM: Semantic Layers Replacing Hub-and-Spoke Master Data
2026 · OWL / SHACL / SKOS, data-mesh federation

Full index — 25+ articles →