Skip to content
View amabito's full-sized avatar
1 follower · 0 following

Achievements

Achievement: YOLOAchievement: Pull Sharkx2Achievement: Quickdraw

Achievements

Achievement: YOLOAchievement: Pull Sharkx2Achievement: Quickdraw

Block or report amabito

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.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

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

Report abuse
amabito/README.md

Runtime enforcement for LLM agents

Building VERONICA — a runtime safety layer that sits between autonomous agents and your API bill.

One unguarded loop on a Friday can mean $12K by Monday.
VERONICA enforces hard limits before damage happens.

→ https://github.com/amabito/veronica-core

What I build

VERONICA (LLM Runtime Enforcement)

  • Hard budget limits (HALT, not alert)
  • Step limits & loop control
  • Concurrency gating
  • Crash-resilient state transitions
  • Designed for engineers, not dashboards

HyperRasterizer (3D Gaussian Splatting)

  • Custom CUDA rasterizer (Apache 2.0)
  • RTX 5090 / Blackwell optimized
  • Solves commercial licensing constraints in 3DGS

HyperViewer

  • WebGPU-based 3DGS viewer
  • Browser-native, no plugin

Technical Background

GPU / Parallel Systems:

  • CUDA (sm_89, sm_120 / Blackwell)
  • PyTorch CUDA Extensions
  • WebGPU

Languages:

  • Python, C++, CUDA C
  • TypeScript, C#

Focus Areas:

  • Runtime safety for AI systems
  • 3D Gaussian Splatting
  • Custom rasterization pipelines
  • Real-time rendering optimization

Writing

  • Zenn (65 articles) — 3DGS / CUDA / WebGPU
  • Qiita (6 articles) — engineering notes

Background

Construction consultant × GPU engineer.

Working at the intersection of infrastructure DX and high-performance graphics systems,
now building runtime enforcement for AI agents.

GitHub: @amabito

Pinned Loading

  1. veronica-core veronica-core Public

    Zero-dep runtime enforcement for LLM agents. Budget limits, concurrency gates, degradation control.

    Python 1