CarvedCoder

ML Systems Engineer | C/C++ | Performance Engineering

Building machine learning systems from first principles—neural networks in pure C++, tensor operations with manual memory layout, and inference pipelines optimized at the instruction level. I focus on understanding what happens beneath the abstractions.

Technical Focus

Systems Programming · Writing performance-critical code in C and C++. Memory management, cache-aware algorithms, and low-level optimization.

Machine Learning Foundations · Implementing ML algorithms from scratch to understand gradients, backpropagation, and numerical computation—not just calling framework APIs.

Data Structures & Algorithms · Building strong algorithmic foundations. 100+ problems solved on LeetCode with focus on complexity analysis and optimal solutions.

Selected Projects

Neural Network Engine (C++)

A fully-connected neural network library built without external ML frameworks. Implements forward propagation, backpropagation, and gradient descent from mathematical foundations. Manual memory management with efficient matrix operations.

Tensor Library

Custom tensor data structure supporting n-dimensional arrays with:

Stride-based memory layout for row-major and column-major access
Broadcasting semantics for element-wise operations
Memory-efficient views without data copying

Algorithm Implementations

Production-quality implementations of core data structures and algorithms in C/C++. Focus on correctness, edge case handling, and performance characteristics.

Skills

Category	Technologies
Languages	C, C++, Python
Systems	Linux, CMake, GDB, Valgrind
ML/Numerical	NumPy, PyTorch (for validation), manual implementations
Tools	Git, Docker, Make

Learning Philosophy

I believe in understanding systems at their lowest useful level of abstraction. When learning ML, I implement the algorithms in C++ before using PyTorch. When optimizing code, I profile first and understand the hardware constraints.

This approach takes longer initially, but produces engineers who can debug production issues, optimize bottlenecks, and build systems that don't exist yet.

Current focus areas:

Advanced C++ patterns (RAII, move semantics, template metaprogramming)
Computer architecture (cache hierarchies, SIMD, memory models)
Numerical computing and floating-point precision

Career Direction

Working toward ML Systems and Infrastructure Engineering roles at top-tier technology companies. I'm drawn to problems at the intersection of machine learning and systems programming—inference optimization, ML compilers, high-performance training infrastructure.

I want to work on systems where correctness matters, performance is measured, and the engineering is rigorous.

What I Care About in Engineering

Correctness over cleverness. Code that works in all cases beats code that's impressive in the common case.

Understanding over usage. Knowing why something works enables debugging when it doesn't.

Fundamentals over frameworks. Frameworks change; algorithms, data structures, and systems principles don't.