Build a real rotary positional embedding kernel in PyTorch with a custom autograd node, fused CUDA-style kernels, and a tiny test harness. The project you ship on a CV to signal GPU, systems, and ML-engineering depth.
Most ‘AI engineering’ tutorials stop at a notebook. This post walks through the full production stack — data pipelines, training loops, evaluation harnesses, serving, and observability — so you can build a system that actually ships.
Build a streaming token sampler from scratch in PyTorch covering temperature, top-k, top-p, min-p, and contrastive search — a CV-worthy systems project for ML engineers.
Build a working LoRA fine-tuning loop in pure NumPy, computing every gradient by hand through low-rank adapters. A practical side project that demonstrates real systems understanding to hiring managers.
A from-scratch, runnable Python project that mimics the memory management at the heart of modern LLMs: sliding-window attention with dynamic eviction and a prefix-cache KV store. Includes architecture, code, tests, and a roadmap to senior-level.