Systems Architect | Performance Engineer | Electrical Engineer
Specializing in Deterministic C++ Runtimes & Green Computing
- 📝 Technical Writing: EDBCBlog on Medium
- 🚀 Lead Dev: The Nodepp Project — High-performance Async C++.
- ⚡ Performance: My frameworks hit 10K+ RPS (Requests per Second) in a low-power dual-core hardware through SSO and Zero-Copy optimizations.
I am an Electrical Engineer transitioned into High-Performance Software Architecture. My work focuses on Silicon-Logic Parity — the idea that software should respect the physical limits of the hardware it runs on.
Through the Nodepp Project, I have demonstrated that we can achieve 11,000x Virtual Memory efficiency and 13x Infrastructure ROI by moving away from heavy managed runtimes toward deterministic, vertically integrated C++ architectures.
- Green Computing at the Edge: Scaling Resource-Dense Applications through Deterministic RAII and Silicon-Logic Parity.
- Mastering APIfy: A Routing Protocol for Structured C++ Messaging
- Surviving the Shadows: Building an Unbreakable Server on the Dark Web
| Category | Technologies |
|---|---|
| Low-Level I/O | io_uring, epoll, kqueue, IOCP, Linux/Windows Kernel Primitives |
| Memory Systems | Deterministic RAII, Zero-Copy Slicing, Cache-Locality Optimization |
| Concurrency | Stackless Coroutines, Actor-Reactor Models, Async Event-Loops |
| Platforms | Bare-Metal (ESP32/AVR), Linux Servers, WASM, Windows |
A functional implementation of the Simple-As-Possible (SAP-1) computer architecture from Albert Paul Malvino’s Digital Computer Electronics. This project represents my deep understanding of instruction sets, bus architectures, and CPU logic.
A "from-scratch" Image Processing and Machine Learning library for the ESP32 and OV7670.
- Custom Tensor Engine: Pointer-arithmetic math operators and manual memory management designed for the constraints of embedded systems.
- Embedded Vision: Real-time implementation of Haar Cascades, Sobel filters, LBP, and template matching.
- Neural Framework: Built-in MLP (Multi-Layer Perceptron) featuring manual backpropagation and custom activation kernels (ReLU, Sigmoid, Softmax).
Nodepp is my research into the Efficiency Paradox. It is a unified asynchronous C++ runtime that scales from 8-bit microcontrollers to multi-core cloud servers with one codebase.
- 📈 Performance: 2.9MB RSS @ 6,000+ RPS (Outperforming Go and Bun in the same machine).
- 💡 Innovation: Architected
ptr_t(Deterministic Memory),kernel_t(Unified Reactor) andcoroutine_t(stackless coroutine). - 🌱 Impact: Built for Green Computing — reducing data center energy consumption through instruction-per-watt efficiency.
| Project | Description |
|---|---|
| Clustor | A high-availability Tor Hidden Service load balancer built with Nodepp. |
| Termia | An asynchronous terminal-based AI assistant optimized for low-latency interaction. |
| Stable Diffusion Raylib | AI image generation integrated directly with Raylib. |
| Arduino Enigma | Asynchronous Enigma Machine simulation running on an Arduino Nano (2KB RAM). |
| Duck Hunt VR | High-performance WASM port featuring AR marker tracking and low-latency orientation. |
| Raylib Gemini Chat | Real-time LLM integration within a native C++ rendering engine. |
- Email: becerracenmanueld@gmail.com
- Ko-fi: Support my open-source research
- Medium: Deep-dives into C++ and Hardware Architecture
"Software is just an arrangement of electrons. I make sure they're moving at maximum efficiency."