Bowen Zhang

张博文

Physics & Computer Science double major at Vanderbilt, with industry experience in ML engineering and a research background in computational physics.

Open to research opportunities

I'm a Physics & Computer Science double major transferring to Vanderbilt University this fall, with industry experience building ML-powered enterprise tools at Lenovo and a research background in computational hydraulics.

My interests sit at the intersection of physics and computing — from Kalman filters on raw sensor data to RAG-based LLM platforms in production. I'm drawn to problems where first-principles modeling meets real-world data.

Aug. 2026 — 2029

Vanderbilt University

B.S. in Physics & Computer Science (Double Major)

College of Arts & Science · Transfer · Expected May 2029

2025 — 2026

UC Santa Barbara

Undergraduate Studies

Mechanics · Fluid Mechanics & Thermodynamics · E&M · Multivariable Calculus · Linear Algebra · Differential Equations · Data Structures & OOP (C++)

Jun. 2026 — Now

Lenovo Group

Machine Learning Intern · Beijing HQ

  • Extended an enterprise LLM quality-analysis platform built on LangChain-Chatchat: integrated the web frontend with the RAG backend — API proxy, JWT auth, knowledge-base upload, streaming chat, and session restore.
  • Redesigned the analysis workflow from tool-driven menus to a problem-driven investigation flow across SPC, statistical, and ML modules; fixed data ingestion for real engineering spreadsheets.
2023 — 2025

Nanjing Hydraulic Research Institute

Research Assistant Intern

  • Preprocessed large-scale flume experiment datasets for ANN modeling pipelines; validated simulation outputs against physical measurements.

Computational Physics

Motion Tracker

Discrete Kalman filter built from first principles to estimate position and velocity from noisy pendulum video. Achieved ~90% jitter reduction, validated by recovering g ≈ 9.8 m/s². Presented to Prof. Clifford V. Johnson (UCSB) and industry engineers at CES 2026.

Python OpenCV Kalman Filter
View on GitHub

★ Best Data Visualization · UCSB Datathon 2026

PDE Epidemic Simulation

Infrastructure lead for a high-performance Rust environment running nonlinear PDE solvers, plus an API data-ingestion pipeline for COVID-19 spread modeling.

Rust PDE Solvers Data Viz

Agent

Spongent

An intelligent agent that discovers and recommends videos based on user preferences and content analysis.

Go
View on GitHub
Best Data Visualization — UCSB Datathon 2026
National Second Prize — National Youth UAV Competition (CSAA, 2022), flight systems integration

Languages

  • Python (NumPy, OpenCV)
  • C++
  • Rust
  • Go
  • MATLAB
  • SQL

AI / ML

  • RAG Pipelines
  • LangChain-Chatchat
  • Kalman Filtering
  • ANN Modeling
  • PDE Solvers

Tools

  • Git / GitHub
  • Linux
  • LaTeX
  • Docker

Interested in research collaborations, ML engineering roles, or just connecting — feel free to reach out.