Articles on #LLMs

Jul 23, 202611 min read

DeepSeek-R1 Architecture: How Mixture of Experts (MoE) & Reasoning Fine-Tuning Scale Intelligence

Unpacking DeepSeek-R1's Multi-head Latent Attention (MLA), DeepSeekMoE 210B sparse routing, and group relative policy optimization (GRPO) reinforcement learning.

Sarah Chen

Transformer Attention Mechanisms: Scaled Dot-Product vs Multi-Query Attention (MQA)

AI & LLMs

Jul 22, 20269 min read

Transformer Attention Mechanisms: Scaled Dot-Product vs Multi-Query Attention (MQA)

How Grouped-Query Attention (GQA) and Multi-Query Attention (MQA) reduce KV-cache memory bandwidth consumption in LLM inference.

Sarah Chen

RAG Cache Strategies: Semantic Caching with Redis and Sentence Transformers

AI & LLMs

Jul 21, 20268 min read

RAG Cache Strategies: Semantic Caching with Redis and Sentence Transformers

Avoid repeating $0.03 LLM queries by matching semantically similar user prompts in vector cache before hitting API endpoints.

Sarah Chen

Agentic RAG with Router Agents and Dynamic Tool Calling

AI & LLMs

Jul 20, 20269 min read

Agentic RAG with Router Agents and Dynamic Tool Calling

Moving beyond naive vector lookup: building adaptive multi-agent retrieval routers that decide when to query vector databases vs SQL databases.

Sarah Chen

Direct Preference Optimization (DPO) vs RLHF: Simplifying Alignment for LLMs

AI & LLMs

Jul 19, 20269 min read

Direct Preference Optimization (DPO) vs RLHF: Simplifying Alignment for LLMs

DPO eliminates the need for training separate reward models and complex PPO reinforcement loops by deriving the implicit reward directly from binary preference data.

Sarah Chen

Prompt Security: Defending Against Indirect Prompt Injection in Production LLM Apps

AI & LLMs

Jul 18, 20268 min read

Prompt Security: Defending Against Indirect Prompt Injection in Production LLM Apps

How malicious instructions hidden inside scraped web pages or user uploaded PDFs can hijack your LLM agent's execution flow.

Sarah Chen