https://tamnp.me/I am a freshman at University of Science, VNU-HCM, and having a strong passion for information technology. I created this blog to share my learning journey, personal projects, and interesting stories surrounding the field — especially those related to AI. 2026-06-18T14:45:59+00:00 Phuc Tam Nguyen https://tamnp.me/ Jekyll © 2026 Phuc Tam Nguyen /assets/img/favicons/favicon.ico /assets/img/favicons/favicon-96x96.png 2026-06-16T14:00:00+00:00 2026-06-16T14:00:00+00:00 https://tamnp.me/posts/data-compress/ Phuc Tam Nguyen

Imagine you need to send a 50 MB file to a friend, but your connection is painfully slow. Or picture a server storing millions of images - without compression, costs would be astronomical. Every time you download a ZIP file, stream a YouTube video, or send a photo on Messenger, data compression algorithms are silently working behind the scenes to make it all possible. In this post, I’ll walk th...

2026-06-02T14:00:00+00:00 2026-06-02T14:00:00+00:00 https://tamnp.me/posts/hash/ Phuc Tam Nguyen

In our journey through data structures, we have encountered various paradigms for storing and retrieving memory. Linear structures like standard Arrays provide contiguous memory layouts for lightning-fast indexing, but the degrade to a sluggish $O(n)$ linear scan when searching for a specific key. Linked Lists solve the problem of dynamic allocation but force the CPU into inefficient, $O(n)$ po...

2026-05-29T14:00:00+00:00 2026-05-29T16:15:50+00:00 https://tamnp.me/posts/btree/ Phuc Tam Nguyen

In our previous deep dive into the m-Way Search Tree, we discovered how increasing the order $m$ flattens the tree, reducing its height. However, the pure m-Way Tree has a fatal algorithmic flaw: it lacks a self-balancing mechanism. In the worst-case scenario of sequential insertions, it degrades into a linear linked list, resulting in $O(n)$ access time. To power robust, large-scale systems l...

2026-05-23T14:00:00+00:00 2026-05-29T16:15:50+00:00 https://tamnp.me/posts/mway-tree/ Phuc Tam Nguyen

In previous explorations of data structures, we heavily relied on the binary paradigm - where a node makes a strict binary decision (left or right). However, as data volumes scale, especially in disk-based storage systems, the height of a binary search tree becomes a bottleneck. To minimize the number of memory accesses, we must flatten the tree. This bring us to the generalization of the binar...

2026-05-22T05:00:00+00:00 2026-05-23T09:58:03+00:00 https://tamnp.me/posts/ml-workflow/ Phuc Tam Nguyen

In the pursuit of building intelligent systems, it is a common misconception to view Machine Learning (ML) purely as the act of training algorithms. In reality, model training is but a single cog in a much larger, continuous engineering machine. To ensure models are scalable, reliable, and practically useful, we must adhere to a structured sequence of operations known as the Machine Learning Li...