- Cleveland, OH
- https://jackvial.com
- @jackvial89
Highlights
- Pro
Lists (4)
Sort Name ascending (A-Z)
- All languages
- Alloy
- Assembly
- Astro
- Bikeshed
- Blade
- C
- C#
- C++
- CSS
- Clojure
- CoffeeScript
- Cuda
- Cython
- DIGITAL Command Language
- Dafny
- Dart
- Dockerfile
- Elm
- Erlang
- F#
- GAP
- Go
- Groovy
- HCL
- HTML
- Haskell
- Java
- JavaScript
- Julia
- Jupyter Notebook
- Kotlin
- Lean
- Lua
- MATLAB
- MDX
- MLIR
- Mojo
- Mustache
- OCaml
- Objective-C
- Objective-C++
- OpenSCAD
- PHP
- Pascal
- Perl
- Pug
- Python
- R
- Rich Text Format
- Ruby
- Rust
- SCSS
- SMT
- Scala
- Scheme
- Shell
- Slash
- Standard ML
- Swift
- SystemVerilog
- TLA
- TeX
- TypeScript
- Vim Script
- Vim Snippet
- Vue
Starred repositories
PyBot: a chatbot written in pure Python by a swarm of agents. No weights, no training, just rules.
A multi core friendly rigid body physics and collision detection library. Written in C++. Suitable for games and VR applications. Used by Horizon Forbidden West and Death Stranding 2.
A lightweight, scaleable rigid-body physics solver
Training Reasoning Models in Simulation to Natively Use Real Robots
verification toolchain for TypeScript (Tech Preview)
Bend 2: a fast language that blocks AI mistakes via proof. Install: curl -fsSL https://bend-lang.com/install.sh | sh
Formalizing CPython 3.15.2 in Lean 4; bugs found along the way
GPU-optimized version of the MuJoCo physics simulator, designed for NVIDIA hardware.
Open source evals for physical AI. Run any LLM/VLA on any arm/humanoid against any real/sim benchmark.
GraspSampler is a library that efficiently generates grasping datasets by quickly sampling millions of parallel-jaw grasps around specific target objects.
Exact convex collision geometry for MuJoCo assemblies: pegs and holes from any outline, threads, and gears.
Yosecula: a Dracula-inspired theme for Omarchy
Build calibrated AI Functions from human feedback using Jev and GEPA.
Identify and simulate extended friction models for servo-actuators
Differentiable Finite Element Method with JAX
RL training environments for Microduck (mjlab)
A Python framework for GPU-accelerated simulation, robotics, and machine learning.
My solution for the LeHome challenge (1st place online, 2nd place in real world round)
A GPU-Parallel analytical physics engine with mujoco API compatibility
[RSS 2026] Code for RISE: Self-Improving Robot Policy with Compositional World Model
[ECCV 2026] VLA-JEPA: Enhancing Vision-Language-Action Model with Latent World Model






