About this site

Portfolio & freelance sandbox — Custom LLMs • Vision AI • Bitcoin & Lightning

Contact Login

What is vinha.ai?

This website is a live portfolio & freelance showcase. It’s where you can explore how I build custom Large Language Models (LLMs), computer vision pipelines, and Bitcoin/Lightning integrations — the same building blocks I use for clients.

  • Demo first: Try the chat bubble (bottom-right) to feel the UX and flow.
  • Vision sandbox: Log in to access image recognition demos and workflows.
  • Engineering focus: Clean architecture, privacy, deploy-ready components.

Try the LLM chat

Open the chat widget (bottom-right). In production I connect it to a private LLM tuned for your domain (RAG, tool use, guardrails). Here, it’s a visual demo to show UX and flow.

Talk about a custom LLM

Test image recognition

After you log in, I can enable a demo page that processes images (detection/segmentation), supporting edge/embedded targets like ESP32, NVIDIA Jetson and Bouffalo MCUs.

Login Create account

Who am I?

I’m Fernando Henrique Vinha, a software engineer specialized in AI/ML, computer vision, embedded systems, and full-stack development. My experience bridges research, product development, and industrial automation — creating real-world solutions that integrate artificial intelligence, IoT, and scalable backends.

Professional Background

  • AI-driven agricultural monitoring and pest detection systems.
  • ERP customization and integrations with Odoo and SAP.
  • Custom Flutter apps for data capture, geolocation, and computer vision.
  • Embedded systems for ESP32, NVIDIA Jetson, and Bouffalo MCUs.
  • Bitcoin Core & Lightning Network payment solutions.

Areas of Expertise

Machine learning (PyTorch, TensorFlow, YOLO, ONNX), computer vision pipelines, backend APIs (Django, FastAPI), mobile (Flutter), ERP customizations, IoT device programming, and payment automation.

What makes my work different?

I combine R&D mindset with production-ready engineering. Whether it's deploying a model to the cloud, integrating it into an ERP, or running it on a microcontroller — the focus is always on delivering a complete, usable, and maintainable solution.

Stack tooling
PyTorch TensorFlow YOLO ONNX ESP32 NVIDIA Jetson Bouffalo MCU Django FastAPI Flutter Odoo SAP CDI Bitcoin Core LND

Work with me

1. Discovery

Quick call to map goals, data, constraints and success criteria.

2. Prototype

Proof-of-value to validate model quality and UX early.

3. Delivery

Production, monitoring and hand-off — cloud, on-prem or edge.

Tell me about your project Read FAQs

FAQ

Can you work under NDA and with private data?
Yes. I can sign NDAs and adapt to your security requirements (network, storage, access). Data stays under your control.
Cloud, on-prem or edge?
Any of them. I design for the target environment, from cloud workloads to on-prem appliances and microcontroller edge deployments.
How do we start?
Kick off with a short discovery call, define success metrics, then a scoped prototype to de-risk. From there we scale.