Hello! I'm Akhlak Mahmood

Product Innovation Manager at Matmerize, Inc.

I am a materials scientist in polymer informatics and AI-accelerated materials discovery. I deliver new-to-the-world polymers for industrial clients in packaging, coatings, fuel cells, batteries, and capacitive energy storage, from property targets to experimentally validated candidates. Complementary expertise in all-atom MD for magnetic nanoparticles and colloidal systems.

Experience

  • Product Innovation Manager
    Matmerize, Inc.
    Atlanta, GA
    • PFAS-free fluoropolymer replacements: designed a multi-tier, halogen-free screening pipeline against a target multi-property envelope; delivered reaction-template-anchored candidate lists with synthetic-accessibility scoring.
    • Bespoke membrane polymer design for industrial clients: encoded client-specific multi-property performance envelopes, ran generative homopolymer screens, and hand-designed copolymer architectures with composition-vs-property analyses for client R&D teams.
    • Proton-exchange membranes for green hydrogen: built an uncertainty-aware screening pipeline across the full PEM performance envelope; validated against industry-standard membranes and delivered halogen-free, synthesizable candidates. Co-author on the resulting manuscript.
    • Formulation → performance ML for industrial consumer products: curated heterogeneous formulation, supplier, and lab data with physics-aware rheology handling; built a zero-leakage nested-CV framework with interpretability panels and a benchmarking CLI used by the client's team.
    • ASKPOLY (sole developer, public launch 2024): Matmerize's LLM-based conversational polymer expert, fine-tuned models over polymer literature and internal knowledge base; enabled tenant-isolated customer fine-tuning.
  • Postdoctoral Fellow, Ramprasad Group, MSE
    Georgia Institute of Technology
    Atlanta, GA
    • Built an NLP pipeline over ~2.4M journal articles combining named-entity recognition and LLM-based extractors; produced a literature-scale dataset across 24 polymer properties, now backing polymerscholar.org and Matmerize's downstream property models (Comm. Materials 2024).
    • Developed a classical-MD protocol for thermoset polymer systems and glass-transition temperature measurement, supporting the group's additive-manufacturing program.
    • Co-authored studies on LLM-based polymer solubility prediction (ACS Materials Letters 2025), LLM benchmarking for polymer properties (Macromol. Rapid Commun. 2025), organic-solar-cell design (npj Comp. Mater. 2025), and PET-replacement copolymer design (Phys. Rev. Materials 2026).
  • Graduate Research Assistant, Yingling Group, MSE
    NC State University
    Raleigh, NC
    • Developed the first all-atom MD method for magnetic nanoparticles with explicit Zeeman alignment and dipolar interactions; shipped as the open-source lammps-mspin LAMMPS plugin (J. Chem. Theory Comput. 2022).
    • Characterized solvent-driven ligand stripping and ligand exchange on colloidal nanoparticles through all-atom MD; mapped solvent-chemistry effects and agglomeration inhibition (ACS Nano 2023; Adv. Mater. Interfaces 2025).
    • Applied ML on small heterogeneous experimental data, Gaussian-process regression, genetic algorithms, active learning, data imputation, to optimize silica-shell synthesis on gold nanorods (Chem. Mater. 2024) and groundwater-contaminant risk analysis (Env. Sci. Technol. 2024).

Education

  • Ph.D. in Physics
    NC State University
    Raleigh, NC
    Graduate Certification in Materials Informatics.
  • M.S. in Physics
    NC State University
    Raleigh, NC
  • B.S. in Physics
    University of Dhaka
    Dhaka, Bangladesh

Skills

Materials Design and Informatics

Multi-property polymer design under entangled trade-offs; multi-task / multi-target co-learning; active learning on small data sets; generative polymer design under regulatory (TSCA) and synthetic-accessibility constraints; formulation optimization; reaction-template-based candidate generation (RxnChainer); polymer fingerprinting (Polymer Genome, PolyBERT, PolyGNN).

Application Domains

Polymer dielectrics and capacitive energy storage; proton-exchange membranes (PEMs) and solid polymer electrolytes (SPEs) for fuel cells and Li-ion batteries; sustainable and biodegradable polymers (PET replacements, PHAs); industrial coatings and resin formulations; nonwovens; nanoparticle self-assembly and colloidal nanomaterials; soft-matter and bio-organic interfaces.

Simulation

All-atom and coarse-grained molecular dynamics (AMBER, LAMMPS, including plugin development); classical force-field development; DFT-based property generation; RDKit; in-silico characterization of mechanical, thermal, optical, and electronic properties.

AI and Machine Learning

Large language models (domain fine-tuning, RAG, agentic systems); transformer and BERT models; graph neural networks; Gaussian-process regression; genetic algorithms; Bayesian optimization; data imputation; literature-scale NLP extraction. Production delivery in Python, PostgreSQL, and AWS.

Selected Publications

Invited Industry Talks

ACS Spring 2026

Atlanta, GA

Polymer and formulation informatics: Achievements and strategic next steps.

ACS Spring 2026

Atlanta, GA

AI-driven polymers and formulations innovations at the industrial scale.

ACS Spring 2025

San Diego, CA

Automating Polymer Design and Chemical Discovery using Generative Modeling.

IDEA25 / FiltXPO 2025

Miami Beach, FL

Unlocking the Future of Nonwovens: AI-Driven Polymer Informatics for Sustainable Innovation.

MRS Fall 2024

Boston, MA

Design of Functional and Sustainable Polymers Assisted by AI.