Blog Metadata Implementation Plan

For agentic workers: REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (- [ ]) syntax for tracking.

Goal: Add tags, subtitles, author, and social-share front matter to all 28 posts, and publish the untracked LLM architectures post.

Architecture: Each task edits a batch of posts by adding missing YAML front matter fields. No content is changed — only the --- block at the top of each file. The untracked post + images are committed in the final task.

Tech Stack: Jekyll front matter (YAML), git

Front Matter Pattern

Every post should end up with this structure (add only the missing fields, don’t duplicate existing ones):

---
layout: post
title: "..."
subtitle: "..."
author: hoangbm
tags: [tag1, tag2]
social-share: true
---

Posts that already have author: hoangbm don’t need it added again. Check before editing.

Task 1: 2018 posts (14 files)

Files: All _posts/2018-*.md

  • Step 1: Add front matter to each 2018 post

    For each file below, open it and add the missing fields to the --- block. Do not change any content below the front matter.

    File Add subtitle: Add tags: Notes
    2018-01-04-first-post.md Why I started writing about machine learning (none) No tags for meta post
    2018-01-08-convnet-element.md The building blocks of convolutional neural networks [deep-learning]  
    2018-01-12-convnet-architecture.md A tour of landmark CNN architectures from AlexNet to ResNet [deep-learning, computer-vision]  
    2018-01-16-challenges.md Common pitfalls when training deep neural networks [deep-learning]  
    2018-01-18-distributed-tensorflow.md Setting up distributed training with TensorFlow [deep-learning, infra]  
    2018-02-25-improvement-distributed-training.md Techniques for faster and more stable distributed training [deep-learning, infra]  
    2018-03-20-image-similarity.md Learning visual representations for similarity search [computer-vision, deep-learning]  
    2018-06-01-sagemaker-tut.md Training and deploying models with Amazon SageMaker [aws, systems]  
    2018-10-20-multi-armed-bandits.md Balancing exploration and exploitation in recommendation [recsys, statistics]  
    2018-10-28-randomization.md Why randomization matters for valid experiment design [statistics]  
    2018-10-30-apache-hadoop-introduction.md A practical introduction to the Hadoop ecosystem [infra]  
    2018-10-8-ner-tagger.md Building a named entity recognition system from scratch [nlp, deep-learning]  
    2018-11-07-hdfs-1.md How the Hadoop Distributed File System stores data [infra]  
    2018-11-10-hdfs-2.md HDFS architecture: NameNode, DataNode, and replication [infra]  

    Also add author: hoangbm and social-share: true to every file that is missing them.

    Example — 2018-01-08-convnet-element.md before:

    ---
layout: post
title: ConvNet and its principal elements
---

    After:

    ---
layout: post
title: ConvNet and its principal elements
subtitle: The building blocks of convolutional neural networks
author: hoangbm
tags: [deep-learning]
social-share: true
---

  • Step 2: Build to verify no front matter errors

    cd /Users/harbui/axon/axon-ai/minhhoangbui.github.io
bundle exec jekyll build 2>&1 | tail -5

    Expected: done in X seconds. — no YAML errors.

  • Step 3: Commit

    git add _posts/2018-*.md
git commit -m "content: add tags, subtitle, author to 2018 posts"

Task 2: 2019 posts (6 files)

Files: All _posts/2019-*.md

  • Step 1: Add front matter to each 2019 post

    File subtitle: tags:
    2019-01-12-amazon-personalize.md Using Amazon Personalize for production recommendation [aws, recsys]
    2019-01-16-retrieval.md Candidate retrieval in large-scale recommendation systems [nlp, recsys]
    2019-02-22-clustering.md From K-Means to DBSCAN: clustering algorithms compared [statistics, deep-learning]
    2019-03-04-mixture-models.md Gaussian mixture models and the EM algorithm [statistics]
    2019-05-19-triplet.md Learning embeddings with triplet loss [deep-learning, computer-vision]
    2019-09-06-pose-extraction.md Estimating human pose from images [computer-vision, deep-learning]

    Add author: hoangbm and social-share: true to any missing them.

  • Step 2: Build

    bundle exec jekyll build 2>&1 | tail -5

    Expected: no errors.

  • Step 3: Commit

    git add _posts/2019-*.md
git commit -m "content: add tags, subtitle, author to 2019 posts"

Task 3: 2021–2025 tracked posts (7 files)

Files: _posts/2021-*.md, _posts/2022-*.md, _posts/2023-*.md, _posts/2025-*.md

  • Step 1: Add front matter

    File subtitle: tags:
    2021-02-06-yolact.md Real-time instance segmentation with YOLACT [computer-vision, deep-learning]
    2021-02-22-solo.md Instance segmentation without bounding boxes [computer-vision, deep-learning]
    2022-12-24-object-detection.md A personal guide to modern object detection [computer-vision, deep-learning]
    2023-01-01-attention.md Understanding the attention mechanism in transformers [nlp, deep-learning]
    2023-11-4-tts.md How modern text-to-speech systems work [nlp, deep-learning]
    2025-02-12-notes-recommendation-system.md A practical three-stage recommendation system pipeline [recsys, systems]
    2025-02-18-notes-k8s.md Kubernetes concepts for ML engineers [infra, systems]

    Note: 2025-02-12-notes-recommendation-system.md already has author: hoangbm — skip that field for it.

    Add author: hoangbm and social-share: true to any that are missing them.

  • Step 2: Build

    bundle exec jekyll build 2>&1 | tail -5

    Expected: no errors.

  • Step 3: Commit

    git add _posts/2021-*.md _posts/2022-*.md _posts/2023-*.md _posts/2025-*.md
git commit -m "content: add tags, subtitle, author to 2021-2025 posts"

Task 4: Publish LLM post and add metadata

Files:

  • _posts/2026-03-26-llm-architectures.md (currently untracked)

  • images/llm/ (currently untracked)

  • Step 1: Verify untracked files exist

    ls /Users/harbui/axon/axon-ai/minhhoangbui.github.io/_posts/2026-03-26-llm-architectures.md
ls /Users/harbui/axon/axon-ai/minhhoangbui.github.io/images/llm/

    Expected: both exist.

  • Step 2: Add front matter to LLM post

    Current front matter:

    ---
layout: post
title: "Demystifying LLM Architectures: Encoders, Decoders, and Inference"
image: /images/llm/transformer.png
---

    Replace with:

    ---
layout: post
title: "Demystifying LLM Architectures: Encoders, Decoders, and Inference"
subtitle: Encoders, decoders, and how inference actually works
author: hoangbm
tags: [llm, deep-learning]
social-share: true
image: /images/llm/transformer.png
---

  • Step 3: Build

    bundle exec jekyll build 2>&1 | tail -5

    Expected: no errors. The new post should appear in _site/.

  • Step 4: Verify post appears in built site

    ls /Users/harbui/axon/axon-ai/minhhoangbui.github.io/_site/2026-03-26-llm-architectures/

    Expected: index.html present.

  • Step 5: Commit

    git add _posts/2026-03-26-llm-architectures.md images/llm/
git commit -m "content: publish LLM architectures post with metadata"

Verification Checklist

After all tasks:

  • bundle exec jekyll build passes with no YAML errors
  • grep -L "tags:" _posts/*.md returns only 2018-01-04-first-post.md (the intentional no-tags post)
  • grep -L "subtitle:" _posts/*.md returns nothing (all posts have subtitles)
  • grep -L "author:" _posts/*.md returns nothing
  • The tags page at http://localhost:4000/tags shows populated tag sections
  • The homepage card grid shows tag pills on post cards