My Blog
Some texts about stuff that I like
DynamoDB
Hash Tables
NoSQL
AWS
Why is DynamoDB So Blazing Fast? A Deep Dive into Hash Tables and NoSQL Performance
DynamoDB
Hash Tables
NoSQL
AWS
Did already ask yourself why GET methods in DynamoDb are so fast?.
Rust
GenerativeAI
Concurrency
LLMs
Rust for Blazing-Fast LLM Applications
Rust
GenerativeAI
Concurrency
LLMs
Python's great for prototyping and has amazing libraries like Hugging Face's Transformers, but when performance matters, Rust shines.
Parallel Programming
I/O Bound
CPU Bound
Python ThreadPoolExecutor and Concurrency: The Complete Guide
Parallel Programming
I/O Bound
CPU Bound
ThreadPoolExecutor is part of Python's `concurrent.futures` module, which was introduced in Python 3.2. It provides a high-level interface for asynchronously executing callables.
AI Engineering
LLM Accuracy
Needle In a Haystack
Machine Learning
Generative AI
Benchmarks
Uncovering The Needle
AI Engineering
LLM Accuracy
Needle In a Haystack
Machine Learning
Generative AI
Benchmarks
In the LLM world, 'needle in a haystack' refers to the challenge of finding specific, accurate information within the vast amount of data these models have been trained on.
RAG Performance
Accuracy
Prompt Engineering
Benchmarks
How to improve your RAG performance?
RAG Performance
Accuracy
Prompt Engineering
Benchmarks
When building a RAG application, it's very common to face low accuracy at the beginning. To overcome this issue, techniques such as prompt engineering, filtered search, score extractor and more are mandatory to reach performances above 80%.
Context Caching
Gemini
Tokens
Prompt
Context Caching in LLMs
Context Caching
Gemini
Tokens
Prompt
Cache input tokens and consult them for subsequent requests.
Parallel Programming
C#
Cuda
Multi-core
Computing Architecture
Processors
What is Parallel Programming
Parallel Programming
C#
Cuda
Multi-core
Computing Architecture
Processors
Parallel programming is a type of computing architecture where multiple processes run simultaneously. This approach can significantly speed up computations by leveraging multi-core processors.