Portfolio ↗ · LinkedIn ↗ · Email ↗
São José do Rio Preto, Brazil · Open to remote opportunities worldwide
About · AI engineering · Projects · Stack · Education · Contact
I'm Flávio Martil, a full stack developer with 10+ years of experience building web applications, APIs, and mobile products. My current focus is AI agents, Model Context Protocol (MCP), and LLM integrations.
My work spans TypeScript, React, Node.js, PHP, and Python: from multi-tenant SaaS applications to tools that connect coding agents with browsers, persistent memory, and external services.
Explore capabilities in text
Tools that connect AI assistants with browser automation, external services, and application workflows through the Model Context Protocol.
Shared persistent memory between coding agents, with SQLite FTS5, synchronization, and orchestration workflows.
Application integrations with Claude, OpenAI models, Gemini, and AWS Bedrock, including tool access and retrieval workflows.
Video processing and analysis workflows combining FFmpeg, AWS Rekognition, and Gemini for AI-assisted commercial compliance checks.
Student support, sales copilots, and workflow automation, including WhatsApp integrations and n8n.
TensorFlow and LSTM experiments, forecasting, data processing, and automated reporting pipelines.
dev-browser ↗ · Browser automation with MCP.
token-cost-calc ↗ · LLM cost estimates with BRL/USD conversion.
claude-mem-opencode ↗ · Shared memory between coding agents.
claude-code-toast ↗ · Windows notifications for Claude Code in WSL2.
Marketplace · API ↗ · Mobile ↗ · An MBA project using Node.js and React Native.
Video analysis pipeline · Private application for AI-assisted commercial compliance checks.
Project details, technologies & attribution
Browser automation for Claude Code across Opera GX, Chrome, and Edge. Extends Sawyer Hood's dev-browser with a dedicated MCP server, side panel, and tab management.
Built with: TypeScript, CDP, MCP, WebSocket.
A web interface for estimating LLM token and API costs, with BRL/USD conversion.
Built with: Next.js, TypeScript, Tailwind CSS.
Shared persistent memory between Claude Code and OpenCode.
Built with: TypeScript, Bun, SQLite FTS5.
Custom Windows notifications for Claude Code running in WSL2, with a glass-style interface built using native desktop components.
Built with: PowerShell, WinForms, DWM, WSL2.
A REST API and mobile application developed as an MBA project, connecting a Node.js backend with a React Native client.
Built with: Node.js, Prisma, React Native, Expo.
API repository ↗ · Mobile repository ↗
Video processing and AI-assisted commercial compliance checks. Source code is not public.
Built with: Python, FFmpeg, AWS Rekognition.
More projects & private applications
Token-efficiency proxy for Codex and Responses API workflows, with recoverable storage, an MCP interface, optional visual compaction, and effort routing.
Built with: TypeScript, MCP, API proxy.
Framework-agnostic tenant context for Node.js, with an optional Mongoose adapter that scopes queries and rejects scoped operations without a tenant.
Built with: TypeScript, Node.js, Mongoose.
Multi-tenant fiscal document processing in production.
Built with: TypeScript, PHP, Node.js.
AI-assisted student support with vector search and automated tests.
Built with: Python, MCP, pgvector.
Explore the full stack in text
Frontend & mobile
React, Next.js, Astro, TypeScript, JavaScript, Tailwind CSS, HTML, CSS, Sass, React Native, Expo.
Backend
Node.js, NestJS, Express, PHP, Laravel, REST APIs.
Data
PostgreSQL, MySQL, MongoDB, Redis, Prisma, SQLite FTS5, pgvector.
AI & automation
Python, Claude API, OpenAI API, Gemini, AWS Bedrock, MCP, n8n.
ML & video
TensorFlow, LSTM experiments, FFmpeg, AWS Rekognition.
Infrastructure & developer tools
Docker, Kubernetes, AWS, Linux, Git, Postman, VS Code, Bun, WSL2.
Currently studying AI, Data Science & Big Data at PUCRS, Applied AI Engineering at UNIPDS, and Cloud Computing Projects at Descomplica.
Academic background
| Program | Institution | Status |
|---|---|---|
| Degree in Systems Analysis & Development | UNIP | Completed |
| Postgraduate specialization in Full Stack Development | Descomplica | Completed |
| MBA in Full Stack, Mobile & Backend Development | Rocketseat | Completed |
| MBA in AI, Data Science & Big Data | PUCRS | In progress, 2026 to 2027 |
| Postgraduate specialization in Applied AI Engineering | UNIPDS | In progress |
| Postgraduate specialization in Cloud Computing Projects | Descomplica | In progress |
Certifications & Anthropic coursework
Datadog Foundation certification, plus additional training in TypeScript, React, Node.js, NestJS, React Native, and Clean Architecture.
The links below point to course pages, not personal certificate verification pages.
| Course | Curriculum |
|---|---|
| Introduction to Model Context Protocol | Course page |
| Model Context Protocol: Advanced Topics | Course page |
| Building with the Claude API | Course page |
| Claude Code 101 | Course page |
| Claude Code in Action | Course page |
| Introduction to Subagents | Course page |
| Introduction to Agent Skills | Course page |
| Claude with Amazon Bedrock | Course page |
| Claude with Google Cloud Vertex AI | Course page |
| AI Fluency: Framework & Foundations | Course page |
| Claude 101 | Course page |
| Introduction to Claude Cowork | Course page |
| AI Capabilities and Limitations | Course page |
| AI Fluency for Students | Course page |
| AI Fluency for Educators | Course page |
| Teaching AI Fluency | Course page |
| AI Fluency for Nonprofits | Course page |
I'm open to remote roles with international teams, particularly in full stack product development, AI engineering, and developer tooling.