Fabio Galuppo

Explore José Antonio Siqueira de Cerqueira's efforts in enhancing language models for greater safety and reliability. Dive into this article to discover his significant contributions to the field - https://lnkd.in/dA-EWMu2! 📚👨‍💻 #research #languagemodels

4

Look what's in C++23 Labels at the end of compound statements (P2324R2), e.g. void f(int& x) {    if (x)      goto END;   x = 42; END: }

20

41 comentários

Implementing Vanilla and Quanto CDS Pricing in C++

50

💬 Short video interview with Flavio Junqueira, CTO at Qbeast, at #JOTB24. Flavio is a renowned computer scientist and researcher known for his significant contributions to the field of distributed systems. #data #bigdata #datalakehouse #distributedsystems 👇👇👇

10

The summary is almost as good as the paper. Thanks. I like the smooth progression in the description of various models; from Variational auto encoder to Hierarchical VAE and finally, Variational diffusion model. Enlightening!

2

Double VS Float in C++ – The Difference Between Floats and Doubles When we want to store numbers with decimals, we can either use the float or double. Though these two data types are used for a similar purpose, they have some differences. The below article discusses the differences between floats and doubles in C++ along with some examples. Float is more memory efficient, but less precise. The byte size for float is 4 while the byte size for double is 8. This implies that double can store values that are twice the amount that float can hold. link: double vs float https://lnkd.in/exE8DprE #quant #finance #programming #cplusplus #cpp #memory #precision

9

1 comentário

🚀 Exploring the Spectrum: Primitive Data Types in C Language 🚀 The C programming language, developed in the early 1970s by Dennis Ritchie, remains one of the most widely used programming languages. Its influence can be seen in many modern languages, and it is known for its efficiency and control over system resources. One of the foundational aspects of C is its set of primitive data types, which are essential for defining variables and managing memory. Understanding these data types is crucial for anyone learning C, as they form the building blocks for more complex data structures and algorithms. For more details click the link below: 👇 https://lnkd.in/gaJGawfm  #CProgramming #ProgrammingBasics #DataTypes #Coding #SoftwareDevelopment #ComputerScience #ProgrammingTutorial #TechEducation #CodingSkills #LearnToCode #SoftwareEngineering #piestsystems #piembsystech #ProgrammingLanguages #CDeveloper #TechTips #ITSkills #Development #TechTraining #pievcore #ProgrammingFundamentals #BeginnerProgramming #TechLearning

12

7 comentários

I am finally on the kubestronaut Map! 🚀 🇦🇷 https://lnkd.in/dCSJPP9Z As now being one of the first two persons to achieve kubestronaut from Argentina, I think I will address a few questions that I received from people on chat and in-person. 1. What is Kubestronaut and how to achieve it? Kubestronaut is a program created by Cloud Native Computing Foundation (CNCF) and The Linux Foundation to identify a list of people called as Kubestronauts who showed exemplary Kubernetes knowledge by passing all five of CNCF’s Kubernetes-related certifications: ✔️Certified Kubernetes Administrator (CKA) ✔️Certified Kubernetes Application Developer (CKAD) ✔️Certified Kubernetes Security Specialist (CKS) ✔️Kubernetes and Cloud Native Associate (KCNA) ✔️Kubernetes and Cloud Security Associate (KCSA). 2. How can I prepare for these certifications? There are a lot of materials online. For all the certifications except KCSA, KodeKloud already provides a list of curated courses and certifications. While they are still working on bringing the KCSA course, you can study that from the documentations in kubernetes.io and also from the concepts explained in CKS course. There are also a lot of options available online, the reason I prefer KodeKloud is because of the access provided in terms of learning as well as hands-on labs and mock exams. Killer Shell labs are also a great way to prepare for the certifications. 3. What if I have a subset of certifications already avaIlable, do I need to recertify them again and then give the remaining certifications? No. As long as these certifications are valid, you can go ahead and do the remaining certifications. To be recognized as a kubestronaut, you need to have all the 5 certifications active regardless of whenever you give them. 4. Do I need to buy the certifications as a bundle or individually? There is no mandate on how you buy the certifications. Bundle does provide some great discounts, so it will be a monetary benefit for you to buy a bundle. So you can choose to buy the certifications any way you want. Do check this out (https://lnkd.in/dTZWRKPC) for discounts for the certifications. 5. (This was asked specifically in my case) I did all the 5 certifications in just a single month. Should anyone else do the same? Not until you feel confident about doing it. It can be taxing for you to do it in such a short time. Trust me, speaking from experience 😅 . The only reason I did it in such a quick way, is because I had a great prior experience working with Kubernetes and my prior strong knowledge with Infrastructure, Application and Security skills. I would suggest everyone aiming for this to take it easy and take it only when you are confident (and not over-confident) while preparing for these certifications. 👉 Book a session with me: https://lnkd.in/dsVZSCv9 #Kubestronaut #Kubernetes #CloudNative #CloudCertifications #TheLinuxFoundation #Argentina

49

6 comentários

I took a graduate database course at the University of São Paulo 13 years ago where the professor said: "NoSQL databases are all the rage and we'll cover them at the end of the course. But first you need to understand relational databases to know everything you loose when you choose NoSQL: everything the NoSQL database doesn't do for you which you'll have to reinvent, implement, and test in your application. Essential features like transactions updating multiple records reliably and safely in the face of concurrent requests." If you're not at the scale of thousands of servers, it's most likely that a relational database will make your life EASIER and your app MORE RELIABLE than a "cool" NoSQL alternative.

528

42 comentários

🔥 SQLite-Vec == VectorDB + OnDevice + SQLite - small memory footprint - written in C - no dependencies pip install sqlite-vec 🎁 GitHub: https://lnkd.in/gRSSEgJe 📚 Docs: https://lnkd.in/gsxGqFCF 🎉 Blog: https://lnkd.in/gJDUYFJw 👉 What is SQLite-Vec: An extremely small, "fast enough" vector search SQLite extension that runs anywhere! A successor to sqlite-vss #database #sql #vectordb #ml #python #c #pylib #sqlitevec #sqlite #datascience

13

New K4 against RocksDB (yes I bench in c++) write benchmarks shows K4 v1.9.0 16 times faster. Here's the last 2 benches from RocksDB RocksDB PUT: 2.77832 seconds RocksDB PUT: 2.91654 seconds Writing 1 million key value pairs, same configuration. 4mb flush. Vs K4 Put time: 181.363391ms Put time: 131.113331ms That's crazy !! What's interesting about this, K4 was at a similar write speed in beta version, but after many iterations now has optimizations using queues and background processes which makes writes blazing fast. #k4 #storageengine #opensource #lsmtree #algorithms

8

10 comentários

Exploring Finite Fields with Rust: Efficient Modular Arithmetic #rust #rustprogramming #rustlang https://lnkd.in/dEsV3n4C

21

2 comentários

Mappings Unlike Solidity, Solana lacks a built-in mapping data structure. However, we can replicate the key-value mappings functionality in Solana by utilizing the HashMap type from the Rust standard library. Unlike EVM chains, the map we are demonstrating here is in memory, not storage. EVM chains do not have in-memory hash maps.

1

📚 Just came across an fascinating set of lecture notes on "𝐓𝐞𝐧 𝐋𝐞𝐜𝐭𝐮𝐫𝐞𝐬 𝐚𝐧𝐝 𝐅𝐨𝐫𝐭𝐲-𝐓𝐰𝐨 𝐎𝐩𝐞𝐧 𝐏𝐫𝐨𝐛𝐥𝐞𝐦𝐬 𝐢𝐧 𝐭𝐡𝐞 𝐌𝐚𝐭𝐡𝐞𝐦𝐚𝐭𝐢𝐜𝐬 𝐨𝐟 𝐃𝐚𝐭𝐚 𝐒𝐜𝐢𝐞𝐧𝐜𝐞" by Afonso S. Bandeira. 🎩 This comprehensive work presents 42 open problems at the intersection of mathematics and modern data analysis. 🔑 Key highlights: - Deep dives into Principal Component Analysis in high dimensions, including critical insights on when PCA works (and when it doesn't) - Novel perspectives on spectral clustering and graph-based learning methods - Rigorous treatment of concentration inequalities and their applications to random matrices - Cutting-edge material on compressed sensing and sparse recovery - Elegant connections between seemingly disparate areas like coding theory and community detection 👀 What I found most compelling was how the notes bridge pure mathematical concepts with practical data science applications. The author does an excellent job explaining how theoretical results translate to real-world algorithmic improvements. 🧵 The open problems presented range from fundamental questions in random matrix theory to practical challenges in data clustering and signal processing. Many remain unsolved and could lead to breakthrough advances in how we process and analyze data. 🧵 If you're interested in the mathematical foundations of data science, this is definitely worth checking out. It's particularly valuable for those working at the intersection of mathematics, computer science, and data analysis. #Mathematics #DataScience #MachineLearning #Research #OpenProblems 👇👇👇 Im creating a lot of scientific content. You're invited to follow me at: 🎙️ Spotify: - https://lnkd.in/d2Y-ehe7 - https://lnkd.in/d-gMtCrE (Hebrew) 🚇 Youtube: https://lnkd.in/dQNqdvMS ☎️ Telegram: https://lnkd.in/dVVqhNw5 (in Hebrew) 🦜 Twitter: https://lnkd.in/dTse8avN

229

6 comentários