CV

Skills

• Software: Physical simulation, dynamical systems, modular code, version control
• Machine Learning: PyTorch, scikit‑learn, Python (10 years exp.), clustering (k‑means), (non)linear optimization, mut. information
• Statistical/ML Methods: Time series prediction, signal processing, information embedding, PCA, self‑supervised learning, data vis., k‑Nearest Neighbors, stability analysis, Fourier analysis, param. optimization

Work experience

• Feb 2025 – Apr 2025: Ascend OS, startup – AI Scientist, Part Time Contractor
  • Built agentic LLM introduction backend for two interested parties
  • Architected and wrote full LLM workflow for icebreakers and table themes for Large (nearly 100 person) corporate event hosting representatives of many corporations
  • Agentic LLM (tool use) via LangGraph
  • Software: Cloud database, issue ticketing
  • Small team production‑level software development
  • Version control: GitHub
• Jun 2024 – Jul 2025 (1 year): Imagia, startup – Computational Physicist
  • Physical simulation: Fully architected end‑to‑end Python pipelines for electromagnetic wave optical metasurface simulation, streamlining design process
  • Lumerical FDTD, RCWA
  • Deconvolution/deblurring
  • Efficient processing: 300× runtime reduction of simulation setup, (8 hour simulation setup reduced to 103 seconds), eliminating a heavy design bottleneck.
  • Built GUI for users to process experimental data directly from measurements, and post‑processing on validation data to reveal accurate results
  • Hardware and algorithm co‑design
  • Reduce datafile size by 4× using geometry symmetries
  • Automated data tracking for simulation results for easy review/reference
  • Worked with team members spanning a complete design pipeline: Virtual design, physical simulation, communicating with chemical fabrication engineer, experimental testing, and post‑experimental data analysis.
  • Researched potential low‑power neural network hardware implementations on market for post‑processing image data
  • Designed, simulated, & presented production‑ready files to fabrication engineers
  • Data preprocessing: noise cleaning/reduction of experimental data
  • Tested noise sensitivity of model with synthetic data
  • Tested different design strategies: Greedy, evolutionary, Monte Carlo, k‑means clustering
  • Data postprocessing: Linear regression, convex optimization, PyTorch feedforward neural network
  • Matched filter algorithms
  • Managed Ansys Licenses, (de)installation and determined GPU requirements
• Aug 2017 – Mar 2018: Intern Research Assoc., LHC ATLAS and PDG, Lawrence‑Berkeley Nat. Lab
  • C++ large data analysis/processing for LHC ATLAS. Implemented unfamiliar framework (C++ ROOT) in 3‑4 months to deliver new results. Poster presented.
  • Production Python data vis. software, Particle Data Group. Delivered complete results in 3‑4 months. Credited Technical Associate in author list: Review of Particle Physics (2018). Poster presented.

Personal projects

• Project: FDTD from scratch (Yee lattice)
  • ○ Parallelized implementation
• Project: Implement MEEP open‑source optics simulation library with parallel processing
  • ○ Tested on multiple simultaneous CPU cores
• Research: Circuits for Low‑Powered Applications, Team Lead
  • ○ Working with one experimental materials scientist and one computational physicist
  • ○ Literature review, experimental design, and analysis
  • ○ Interdisciplinary project: robotics, physics, and neuroscience
  • ○ Data collection in progress

Education

• UCSD, Doctorate in “Physics, with Specialization in Computational Neuroscience” 2024
• University of Arizona, Bachelor’s in Science “Physics and Applied Mathematics” 2017