Guided reading
Collections
Related notes arranged into deliberate paths. Pick a subject, begin with the first idea, and follow the sequence as the system comes together.
- Paths
- 4
- Notes
- 16
Collection
LLM Inference & Serving
Understand how prompts become tokens, how attention and KV cache work, and how serving engines scale inference.
4-note reading path- Read note
From Prompt to Response: A Step-by-Step Walkthrough of LLM Inference
Update: For a deeper systems-level treatment of LLM inference, especially the interaction between request scheduling, prefill, decode, and KV-cache reuse, see arXiv:2606.24937.
- Read note
Attention Dilution
Attention dilution (also called context dilution) is one of the fundamental limitations of transformer-based LLMs when dealing with long contexts or extended agent memory.
- Read note
KV-Centric LLM Serving: vLLM, SGLang, and Disaggregated Attention
The more I look at LLM serving, the more it feels like the main object is not the request, the model, or even the GPU.
Collection
AI Agents & MCP
Learn how agents use tools, how MCP connects systems, and how multi-agent workflows coordinate.
4-note reading path- Read note
AI Terminology: Agents, Skills, RAG, MCP, and the Layers Beneath the Hype
How many of these terms do you actually recognize?
- Read note
How LLM Tools work
Tools in Large Language Models (LLMs) Tools enable large language models (LLMs) to interact with external systems, APIs, or data sources, extending their capabilities beyond text generation.
- Read note
MCP Transports
Compare MCP transport options, from local standard I/O to streaming HTTP connections.
Collection
Practical LLM Applications
Build local chat applications, RAG pipelines, reranking workflows, and natural-language data tools.
4-note reading path- Read note
Gradio with Ollama
Build a lightweight chat interface for a local Ollama model with Gradio.
- Read note
Autogen Intro and RAG Workflow
AutoGen is a framework for creating multi-agent AI applications that can act autonomously or work alongside humans.
- Read note
RAG-Reranking
Retrieval-Augmented Generation (RAG) is a powerful approach that combines retrieval and generation to produce high-quality responses. However, the quality of the final response can be...
Collection
Spark & Data Engineering
Learn DataFrame operations, SQL tuning, Spark optimization, and structured streaming.
4-note reading path- Read note
Spark Dataframe window function
scala ref create dataframe
- Read note
Spark SQL
Configure Spark SQL and tune distributed queries for more efficient execution.
- Read note
Spark Optimization
PROCESSLOCAL data is in the same JVM as the running code. This is the best locality possible NODELOCAL data is on the same node. Examples might be in HDFS on the same node, or in another...