I'm Eduardo, a software developer and Systems Development student from São Paulo, Brazil.
I'm interested in understanding how software works beyond the surface. I enjoy exploring computer science fundamentals, building tools from scratch, and turning ideas into practical software.
My interests span software engineering, artificial intelligence, systems design, and applied mathematics. I'm particularly curious about how mathematical concepts and computational methods can be used to solve complex problems.
Outside programming, I'm interested in astronomy and mathematics. I've also participated in academic olympiads, earning medals in both mathematics and astronomy.
I'm exploring how AI systems can interact with external tools, coordinate tasks, and automate complex workflows.
My current interests include AI agent architecture, tool integration, shared context, and multi-agent orchestration. I also experiment with workflows involving GitHub, Figma, and other development tools.
I'm particularly interested in understanding how these systems can become more reliable, useful, and autonomous.
I'm interested in the intersection of computer science, mathematics, and artificial intelligence.
Some topics I want to explore further:
- Machine Learning: mathematical foundations, optimization, and model behavior.
- Algorithms: computational complexity, data structures, and algorithm design.
- Applied Mathematics: numerical methods, probability, statistics, and optimization.
- AI Systems: agent architectures, distributed coordination, and computational efficiency.
My goal is to develop a stronger theoretical foundation while building practical implementations and exploring opportunities for collaborative research.
typescript · node.js · sqlite · whatsapp cloud api
Asynchronous logistics & post-sales bot built on the official WhatsApp Cloud API. Implements immediate |
python · autograd · neural networks
Scalar automatic differentiation (autograd) engine and neural network library built from scratch in pure Python with zero external dependencies, tracking dynamic computational DAGs and backpropagation. |
javascript · node.js · sha-1 · zlib
Implementation of Git's internal object model ( |
typescript · react · spiking neural networks
Experimental simulation connecting a biological Drosophila connectome subgraph (FlyWire) and a Leaky Integrate-and-Fire (LIF) spiking neural network engine to a real-time interactive game loop. |
react · express · supabase
A web platform designed to support the screening and management of social-assistance placements in Guarulhos. Built with React, Node.js, Express, and Supabase. |
python · streamlit
A Python and Streamlit dashboard for monitoring delivery delays, comparing carriers, filtering performance metrics, and exploring logistics data. |
View all my GitHub repositories →
tobymao/sqlglot(Python SQL Parser & Transpiler) — Merged PRs including overloaded TeradataTRUNCandDateTruncAST transpilation (#8980), Teradata case-insensitive identifier normalization (#8940), SQLiteDATEDIFFday-unit transpilation (#8914), and AST documentation fixes (#8913).open-telemetry/opentelemetry-python-contrib(CNCF OpenTelemetry Python) — Merged PR#5298ensuring proper disposal of async SQLAlchemy engines in instrumentation test suites.langchain-ai/langchainjs&openbq-org/OpenBB— Contributions addressing runtime context overrides inllmToolSelectorMiddleware(#12202), SEC XBRL taxonomy label parsing (#7831), and target downside deviation for Sortino ratio inopenbb-technical(#7834).
- 🥇 Gold medal — Brazilian Astronomy Olympiad (OBA)
- 🥉 Bronze medal — Brazilian Mathematics Olympiad (OBMEP)
- 🏆 2nd place — Programming competition, competing against teams from FATEC
- 🤝 Hackathons and collaborative projects focused on practical problem-solving
- 💻 Software development for real businesses, including management systems and digital platforms
Building things, understanding how they work, and exploring what's possible.

