Python

Cover image showing INT4 quantization flow: FP16 weights compressed to INT4 with block-wise scales

Build an INT4 Block-Wise Weight-Only Quantizer for Transformer Attention Layers

Build a production-grade INT4 block-wise weight-only quantizer from scratch in pure Python. This guide covers calibration, dequantization, benchmarking, and why this project signals senior ML-systems skill to hiring managers.

Conceptual diagram of a dynamic context buffer

Dynamic Context Buffer with Learned Relevance Predictor

This post walks you through implementing a token‑aware context buffer that learns relevance and prunes low‑value tokens to fit a target window.

A schematic of memory-mapped file access across sharded model weights.

Building a Safetensors Schema Validator and mmap-Backed Tensor Reader for Sharded Model Weights

This project demonstrates how to validate safetensors schemas and read sharded model weights using memory-mapped files, showing practical systems expertise.

Building a GGUF Weight Loader from Scratch in Pure Python

A hands-on guide to building a zero-dependency GGUF weight loader from scratch in Python, covering the binary container format, tensor extraction, and a roadmap to production-grade features.

A conceptual diagram of a RAG pipeline with document ingestion, vector store, retriever, re-ranker, and LLM generator stages.

Building a Production-Grade RAG Pipeline: A Hands-On Portfolio Project That Signals Real Systems Skill

Move beyond the generic FAISS + LangChain + OpenAI tutorial. This guide walks you through building a multi-stage RAG pipeline with hybrid retrieval, cross-encoder re-ranking, and automated evaluation — the kind of project that signals real systems skill on a CV.