About
Damilare Osibanjo
Full-spectrum software engineer based in Lagos, Nigeria — available globally for remote roles. I build across the entire computing stack, from bare-metal kernels to machine-learning systems and modern web platforms.
Engineering approach
Software engineering is complete when mastery spans from the bare-metal kernel and machine-learning algorithms down to high-throughput backend RPCs and considered user interfaces.
I build robust infrastructure across the computing spectrum: bare-metal x86 operating system kernels, compilers targeting C and LLVM IR, PyTorch Graph Neural Network recommenders, high-performance TCP RPC frameworks, and full-stack web applications.
Disciplines
Four pillars of the practice
Mechanical sympathy & kernels
Designing execution pipelines with explicit awareness of CPU cache lines, page tables, SIMD alignment, and memory layout.
AI / ML & algorithmic precision
Implementing deep learning architectures like Graph Neural Networks from first principles, with mathematical rigor.
High-throughput backend infrastructure
Building low-latency networking primitives, RPC frameworks, and async logging engines with zero bloat and a low memory footprint.
Full-stack web architecture
Crafting fast, accessible, responsive interfaces powered by modern TypeScript runtimes, React, and server-rendered web frameworks.
Selected work
Projects across disciplines
Bare-metal x86 operating system kernel in Rust. Implements physical/virtual memory paging, a pre-emptive multi-tasking scheduler, RAMDisk VFS, and VGA terminal drivers.
Graph Neural Network recommendation engine in PyTorch. Implements LightGCN sparse matrix message passing, BPR loss, and self-supervised contrastive learning.
Zero-dependency C99 JSON-RPC 2.0 framework over multiplexed TCP streams with binary framing and ring buffer flow control.
Systems language compilers targeting LLVM IR and C99 code generation. Recursive descent AST parsers, static type checking, and linear ownership borrow models.
Hardware-isolated microVM sandbox runtime in Rust. Uses KVM and Apple's Virtualization.framework for sub-5ms cold-boot untrusted code execution.
Production web applications, interactive dashboards, zero-dependency schema validators, and high-performance frontend interfaces.
Background
Self-directed software engineer with over five years of engineering discipline. Originally trained in Economics, I moved deep across the stack out of a drive to understand computation end to end — from hardware and kernels up through ML models and full-stack web platforms.