Turning complex science, AI, and compute into enterprise solutions.

I design and build AI, scientific computing, and high-performance systems for hard problems in any industry. My work so far spans power generation, pharma, biotech, and real-time conversational AI.

View expertise → Get in touch
AI | Physics-Informed AI | Cloud | Scientific Computing | HPC

From equations to architecture to shipped products.

Audivi AI E-Signal DiPhyx Merck & Co. Institute for Clean & Secure Energy University of Utah

Expertise

Five disciplines, each proven in real work.

Artificial intelligence

  • Real-time conversational AI. As Head of Engineering at Audivi AI, I guide architecture and priorities for a product deployed at major fast-food chains. Visit Audivi AI →
  • ML and NLP in pharma. Pipelines for nonclinical drug safety at Merck, integrated with Databricks in a regulated setting.
  • AI-assisted pipeline design. The DiPhyx platform uses AI to help design pipelines for simulation, analytics, and data processing. Visit DiPhyx →

Physics-informed AI

  • Neural operators for turbulence. Fourier neural operators trained on One-Dimensional Turbulence, testing what a surrogate of a stochastic simulation should learn. Read the ODT study →
  • Digital twin for power plants. I architected and developed the digital twin behind E-Signal’s CondenserSentinel. Physics-based models run on data plants already collect, flagging efficiency drift and ranking likely causes such as air in-leakage or tube fouling. Visit E-Signal →

Cloud

  • Multi-cloud scientific platform. DiPhyx designs, runs, and monitors pipelines across AWS, Azure, GCP, and on-prem HPC, catching failures early. Visit DiPhyx →
  • Workflow engine. dxflow turns accessible machines into one computational fleet, with a CLI, REST API, web UI, and built-in authentication and authorization. Visit dxflow →
  • Cloud-enabled simulation. GPU- and cloud-enabled workflows for multiphysics codes at the Institute for Clean and Secure Energy.

Scientific computing

  • Two-Phase One-Dimensional Turbulence (ODT). My doctoral model at the University of Utah, with an interactive walkthrough of how it cuts computational cost. Explore the ODT model →
  • Combustion and turbulence research. Peer-reviewed work in CFD and thermochemistry, in journals such as Combustion and Flame. See publications on Google Scholar →
  • Radiation modeling. The principles and methods, explained with a teapot and a dragon. Read the radiation post →

High-performance computing

  • GPU acceleration. Cut the cost of particle–cell coupling in a multiphysics CFD code, first with a better algorithm, then by moving it to the GPU. Read the GPU post →
  • Large-scale multiphysics. Built and optimized simulation codes at the Institute for Clean and Secure Energy, with major performance gains.
  • Cluster orchestration. dxflow runs Docker and Slurm workloads from one engine. Read the dxflow docs →

About

Scientist

A Ph.D. chemical engineer with more than 20 years of turning science into production software.

Leader

Led an international team at Merck, built the engineering team and ran a design-partner program at DiPhyx, and lead engineering department at Audivi AI.

Delivery

Simulation Software, Cloud & HPC platform, ML pipelines, digital twin, and a conversational AI product, all built and put to use.

Get in touch

Or email me directly →
LinkedIn Google Scholar GitHub X (Twitter) Medium DiPhyx dxflow.ai