Skip to content
View tabedzki's full-sized avatar

Block or report tabedzki

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
tabedzki/README.md

Hi, I'm Christian 👋

I'm a Research Software Engineer at the Princeton Neuroscience Institute, where I write and maintain the software and infrastructure behind the NIH-funded BRAIN CoGS project. Before that I did a Ph.D. in Chemical Engineering at Penn, writing CUDA/C++ and MPI code for polymer simulations.

What I'm working on

  • 🧠 BRAIN CoGS data pipelines. These are DataJoint pipelines for Neuropixels ephys and two-photon imaging: Kilosort 4 with DREDge motion correction, suite2p running on the cluster's GPUs through Slurm, and NWB export. All of this sits on a few hundred terabytes of storage that I keep in order.
  • 🐭 Lab tools. These are the tools the lab uses every day: a Streamlit app with SSO for scheduling technicians and training rigs, the code that runs the ViRMEn virtual-reality rigs, and a Raspberry Pi GUI for water delivery.
  • 🗺️ PixelMap. It's a browser-based tool for designing Neuropixels channel maps. I handle the packaging, Docker, and hosting at pixelmap.pni.princeton.edu, and co-authored the JOSS paper. DOI
  • 💰 Actual Budget. I'm a core contributor. I built the Budget Analysis report (category and group filters, stacked bars, CSV export), and I review PRs and triage issues.
  • 🔒 Keeping research code maintainable. I add uv packaging, pre-commit hooks, CI, Dependabot, and secret scanning to lab repos that have never had tests.
  • ⚡ GPU ports. I help researchers move CPU code to GPUs. At the Princeton Open Hackathons, the teams I mentored got roughly 100x speedups with Numba and 1000x with CuPy.

I also send fixes upstream to the tools I depend on, including datajoint-python, element-array-ephys, react-spectrum, and phy.

Before Princeton

At Penn I worked with a researcher at Sandia to port my lab's particle-mesh (TILD) routines into LAMMPS. I also restructured our in-house CUDA code so we could keep adding features to it, and taught lab workshops on Rust, CUDA/Thrust, and Git.

Tools I use most

Python · C++ · CUDA · MPI · Slurm · DataJoint · Streamlit · Docker · GitHub Actions · uv · Linux

Outside of work I row, run, and self-host more services than I strictly need (Immich, Vaultwarden, Nextcloud, TrueNAS, and so on). I'm also a member of US-RSE and helped run logistics for US-RSE'25.

📫 LinkedIn · tabedzki@alumni.upenn.edu

@tabedzki's activity is private