Skip to content
View Rasantis's full-sized avatar

Block or report Rasantis

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 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
Rasantis/README.md

Rafael De Santis

Senior AI Engineer — I build computer vision that runs on edge GPUs, LLM agents on top of it, and the product around both.

Portfolio LinkedIn Email

São Paulo, Brazil (UTC−3) · Open to remote roles worldwide · 🇮🇹 Italian (EU) citizen — no visa required for the EU


What I've shipped

System Result
🏭 Industrial CV platform — JBS & Marfrig, the world's two largest meat processors 25,000+ animals counted daily across 4 plants; real-time weight estimation at 98.5% accuracy (95% target)
🛒 ShopGuard AI — retail security platform I founded and built solo, edge to cloud 90% fewer completed thefts in production, 24/7; sub-second YOLO11 + TensorRT on NVIDIA edge
Pix Safety — model & serving optimization (current role) +30% accuracy, +45% faster inference, −37% cloud cost in month one
🤖 Multi-agent decision engine — LangGraph over live ERP data Autonomous decisions with a confidence gate, human-in-the-loop fallback, Langfuse traces
Soccer AI — match intelligence from World Cup 2026 broadcast footage Player/referee/ball tracking, per-player km/h, live possession, homography-projected tactical map

See these systems actually running — real footage, real overlays → Not screenshots: video of the pipelines in production.


Stack

Vision & ML — Python · PyTorch · YOLO11 · NVIDIA TensorRT · OpenCV · GStreamer · CUDA · ONNX LLM & Agents — LangGraph · LangChain · AutoGen · RAG · Langfuse · GPT-4 · Claude · eval pipelines Multimodal — VLMs (GPT-4V, Claude vision, LLaVA, self-hosted Qwen-VL) · LoRA fine-tuning · SAM Product — FastAPI · React + TypeScript · Node.js · PostgreSQL · event-driven microservices Infra — Docker · Kubernetes · GCP · AWS · Oracle Cloud · CI/CD · NVIDIA Jetson · MLOps AI-augmented dev — Claude Code · Codex · Cursor, daily


Now

Senior AI Engineer III @ Pix Force — owning performance and cost for Pix Safety, a production computer vision product, and running company-wide training on AI-augmented development.

Before that: Founder & CTO @ ShopGuard AI (accelerated by Oracle for Startups, Google for Startups — Top 10 AI startups in Brazil, US$100k in Google Cloud credits — and Antler Brasil) · Co-Founder & Tech Lead @ Vision Labs · AI engineer at Pix Force since 2024 · freelancing since 2021.


What I'm looking for

Senior AI/ML engineering roles — remote worldwide. Computer vision, LLM agents, or full-stack AI products. I've founded twice and shipped alone, so I'm comfortable owning a product end to end — and I'd rather do it inside a strong team.

rafa25santis@gmail.com · rasantis.github.io · LinkedIn

Popular repositories Loading

  1. landing-page-plr landing-page-plr Public

    repositório para venda de produto plr de emagrecimento

    JavaScript 2

  2. AYSNC-MULTITHREADING- AYSNC-MULTITHREADING- Public

    Threaded video capture — decoupling capture from processing to keep the pipeline at ~150 FPS

    Python 2

  3. REPOSIT-RIO-IC-AGROTECH REPOSIT-RIO-IC-AGROTECH Public

    C++ 1 2

  4. Auto-GPT Auto-GPT Public

    Forked from Significant-Gravitas/AutoGPT

    An experimental open-source attempt to make GPT-4 fully autonomous.

    Python 1

  5. StoreGuard StoreGuard Public

    Real-Time Shoplifting Detection with Computer Vision

    C# 1

  6. GIT-Exericio GIT-Exericio Public

    Forked from kenjimaeda54/GIT-Exericio

    Praticando GIT

    HTML