Research engineer combining mathematics, simulation, and machine learning to solve complex problems. My journey spans from aerospace and finite element analysis to modern AI systems, with a focus on large language models and retrieval-augmented generation. I enjoy exploring complex systems and building high-performance software that bridges theory with practical applications. Learn more about me
Leading a team of R&D engineers in the development of treatment prediction models, shape optimization methods, unsupervised classification, and model integration. Separately pioneered a custom LLM RAG chat with agentic flows within Align leading to better document ingestion flows, re-ranking, and citation of documents.
Developed custom transformer based machine learning models for treatment prediction and optimization. Created embedding and classification models to encode treatments, patient outcomes.
Open-source project implementing Finite Element Analysis (FEA) using the Rust programming language. Provides a framework for running FEA simulations, leveraging Rust's safety, speed, and modern language features for high performance and reliability.
Designed for both research and practical engineering applications, offering computational techniques for solving complex structural, thermal, and other physics problems by discretizing them into smaller, manageable elements.
Automated and pioneered new simulation and analysis methods for orthodontic treatment planning.
Implemented distributed simulation compute on premise to enable grander statistical use of the simulation platform and data collection for machine learning applications.
Interactive web demo that visualizes the training of a Multi-Layer Perceptron (MLP) neural network in real time, entirely in the browser using WebAssembly for high performance. Core neural network logic implemented in Rust and compiled to WASM, with Vue.js frontend.
Demonstrates 2D binary classification where the model learns to distinguish points inside user-defined regions using configurable architectures, loss functions (MSE/BCE), and complex mask expressions. Users can watch decision boundaries evolve in real time.
Leading a team of R&D engineers in the development of machine learning prediction models tied with simulation to provide patient specific design.
Developed and validated custom machine learning models for biomedical optimization. Pioneered a RAG knowledge base chat system tailored for domain specific and agentic use.
Pioneered new simulation and analysis methods for quantifying the effect of orthodontic appliances in highly patient specific morphology.
Provided new insight and data metrics for previous and future cases, advancing personalized treatment planning.
Master's thesis for the University of Florida which applies the step-boundary method to contact analysis within FEA.
Contributed to novel software OnSimulation with analytical geometric representation for contact identification and analysis of stick-slip problems.
Launched contact analysis add-on using the immersed boundary method within the finite element analysis software OnSimulation.
Implemented novel geometric representation techniques for improved contact identification and analysis.
Extensive understanding of solid mechanics, simulation and the FEA method, including system design, plasticity, and contact analysis.
Completed thesis on contact simulation in IBFEM on a novel method to represent contact without meshes. Also served as Graduate Teaching Assistant for FEA and computational CAD courses.
Browser based 2D finite element software for solid mechanics problems which uses the immersed boundary method.
Demonstrated the feasibility of running complex FEA calculations directly in web browsers, making simulation more accessible.
test
Interactive web application that simulates the dynamics of a ball on a table using intuitive 3D visualizations. Built with Vue.js and powered by the three.js graphics library.
Allows users to explore physical motion and control systems through real-time, browser-based simulations with interactive 3D graphics.
Interactive web-based application that streamlines the process of solving Kepler's Equation, a core concept in orbital mechanics and astrodynamics.
Implements robust numerical solvers using Newton-Raphson method with real-time visualization of orbital paths and parameter relationships. Features modern JavaScript frameworks with plotting libraries for dynamic user experience.
Research within finite element analysis on non-linear frictional contact contributing to the novel software OnSimulation.
Contributed significantly to analytical geometric representation for contact identification and analysis of stick-slip problems.
Coursework includes orbital dynamics, flight dynamics, CFD, Monte Carlo simulation and control theory.
Experience with material testing, machining and additive manufacturing. Graduated Cum Laude with strong foundation in engineering principles.
Computational tool designed to visualize and analyze the phase portraits of dynamical systems, primarily focusing on systems of ordinary differential equations (ODEs).
Plots vector fields, trajectories, equilibrium points, and nullclines, providing valuable insights into the qualitative behavior of two-dimensional dynamical systems. Supports interactive exploration and parameter adjustments for teaching and research applications.
JavaScript library for solving orbital mechanics problems, allowing conversion between orbital elements and position/velocity vectors for two-body orbital systems.
Features interactive demo with 3D visualization, supports standard Keplerian orbital elements, and provides both forward and inverse transformations. Built with Vue.js frontend and published as an NPM package.
Single-handedly developed a robust, automated e-commerce platform now driving multi-million dollar revenue.
This system handles orders, retail orders, shipping, notifications, and ad optimization with fully automating fulfillment down to label selection and printing.
Designed and implemented a real time notification system within a financial advising and onramp software.
First professional software development experience, working with real-time systems and financial technology.
Researcher, Aerospace Engineer, and Software Developer
Hover over timeline items to explore my professional journey
I'm an R&D Engineer and Machine Learning specialist with a passion for solving complex problems at the intersection of engineering, mathematics, and artificial intelligence. My journey spans from aerospace engineering fundamentals to cutting-edge AI applications, always driven by curiosity and the desire to build meaningful solutions.
I develop production-grade AI systems including RAG platforms, neural networks, and ML pipelines. Currently leading ML initiatives at Align Technology, I've pioneered patient-specific prediction models and comprehensive evaluation frameworks that bridge research and practical applications.
My expertise spans finite element analysis, contact simulation, and computational mechanics. I've contributed to novel software like OnSimulation and developed web-based FEA solvers, always focusing on making complex simulation tools more accessible and intuitive.
I bridge digital algorithms with real-world hardware through microcontroller programming, custom PCB design, and autonomous robotics. My projects include SLAM-enabled robots, drone flight controllers, and embedded systems that operate reliably in unpredictable environments.
I work across the full stack with languages from C/C++ and Rust for performance-critical applications to Python for ML and JavaScript/TypeScript for web development. I focus on writing clean, maintainable code that solves real problems elegantly and efficiently.
I believe the most impactful work happens at the intersection of disciplines. By combining aerospace engineering rigor with modern AI capabilities and theoretical knowledge with hands-on implementation, I strive to create solutions that push boundaries while remaining grounded in solid engineering principles.
I enjoy capturing landscapes, animals, plants, while also enjoying capturing industry or mechanisms. photos.johndews.com.
I enjoy day hikes and multi-day camping trips, immersing myself in nature and enjoying geology and ecology. My favorite places include Yosemite, Sequoia National Park, and the Sierra Nevada mountains, and Mount Shasta
I enjoy 3D printing, PCB design, and building custom electronics. I enjoy integrating machine learning, sensors, and mechanics into practical projects and artistic creations.
Beyond work, I enjoy exploring new programming languages, contributing to open source projects, and building tools that solve interesting problems. Often simulating new problems or words; or trying to find a deeper understanding of a concept..