Cazton
Software Development
Austin, Texas 2,508 followers
World class Software development, Consulting, Training and Recruiting.
About us
Cazton specializes in Internationally acclaimed experts in Web and Mobile Development, Big Data, Artificial Intelligence and major cloud transformations for the Enterprise. We have a great team of Google Developer Experts, Microsoft MVPs, Asp.NET Insiders, Web API Advisors, Microsoft Data Guidance Council members just to name a few. Our consultants have advanced computer science degrees who believe in introducing the best patterns and practices and the latest and greatest rapid development technologies. For more information contact info@cazton.com Chander Dhall, CEO of Cazton, is an awarded Microsoft MVP, Google Developer Expert and world-renowned technology leader in architecting and implementing solutions. He's not only rescued software development teams, but also implemented successful projects under tight deadlines and difficult business constraints. His company has a proven track record of not just saving the client millions of dollars, but also providing an expedited delivery time. Chander's team of experts are internationally-acclaimed speakers in top technical conferences in the world. As an Azure Advisor, ASP.NET Insider, Web API Advisor, Cosmos DB Advisor, and an MVP for Visual Studio and Development Technologies, Chander has insight into new releases on technologies used by millions of developers. His critical advice, impeccable vision, futuristic strategy backed with creating and establishing best practices in the industry are some of the reasons for his unparalleled success. At the same time, he's a voracious and highly-respected speaker. Chander is known to elucidate critical and complex concepts while making them easy to understand. He's one man who can talk to anyone from a junior developer, a senior architect all the way to an executive and knows how to speak their language. He is the author of "Scalability Patterns" and "Building Single Page Application Using ASP.NET Core & Angular". He has utilized his amazing business sense having started
- Website
-
https://cazton.com
External link for Cazton
- Industry
- Software Development
- Company size
- 51-200 employees
- Headquarters
- Austin, Texas
- Type
- Privately Held
- Founded
- 2013
- Specialties
- Training, Web, Windows, iPhone, Android, development, .NET Core, Angular, React, Big Data, AI, Machine Learning, Node.js, Apache Spark, Hadoop, Ignite, Scala, Kafka, Docker, Kubernetes, Microservices, Software Architecture, Java, TensorFlow, Python, and Keras
Locations
Employees at Cazton
Updates
-
CodeLlama-70B vs GPT-4
CodeLlama-70B is free and outperforms GPT-4? Our team was ecstatic upon learning that CodeLlama-70B surpasses GPT-4 in human evaluations. We staunchly believe that healthy competition is invaluable not only for customers but also for the tech companies that develop these products. A vibrant and robust open-source community is crucial for the tech industry. That's why Cazton team members have been actively contributing to open-source projects since our inception over a decade ago. However, given the prevalence of misinformation, particularly regarding AI, we abide by a simple principle: "Trust, but verify." This has led us to benchmark internally more than 250 unique models (automated) and over 100 models (through human evaluations). Yes, CodeLlama-70B is free, open-source, and has made notable improvements. However, it doesn't come close to GPT-4. Take a look at a few screenshots and see for yourself. Should we then dismiss this model? Absolutely not. It shows considerable potential and is completely open-source. We eagerly anticipate a more finely tuned version. And indeed, Meta deserves recognition for making the Llama models open-source. It's an enormous undertaking, and building large language models (LLMs) is costly. Let's not overlook the opportunity to thank Meta for their contribution to the open-source community. #gpt4 #ai #tech #microsoft #azureopenai #codellama70b #llama2 #meta #openai
-
Blogged: Everything you need to know about Azure Cosmos DB (including LLM based vector support) Must watch video! https://lnkd.in/gyT5vKEY
Turns out you use Azure Cosmos DB on daily basis, and may not even know it :) A few of us looked at our day and what portion of it is powered by the apps that use Cosmos DB. Turned out its good chunk - here is a summary in 90 seconds :).
-
53-page Whitepaper Simplified Decoding Orca 2 by Microsoft Research: Insights from Cazton #cazton #microsoft #gpt4 #gpt4turbo #llm #openai #azureopenai #microsoftresearch #orca #orca2
-
🌟 Grateful to the Microsoft .NET team and .NET Conf for the incredible opportunity to demonstrate building a custom chatbot in just 30 minutes! 🙌✨ (Video and code shared) 🚀 Building with Azure OpenAI and .NET 8 🛠️ Setup and Prerequisites 🌉 Azure OpenAI vs. OpenAI 🧩 Implementation and Project Structure 🎓 Customizing the Chatbot's Knowledge https://lnkd.in/gMGyxjJR #cazton Cazton #net #dotnet #azureopenai #openai #gpt4 #gpt4turbo #gpt35turbo #llm #microsoft #azure
.NET 8 and Azure OpenAI | Cazton
cazton.com
-
MUST READ: Why AI Projects Fail? 10 Problems & Solutions for Domain-Specific LLMs (for the enterprise) Cazton #cazton #llm #ai #openai #azureopenai #bard #claude #llama2 #rag #domainspecificllms #enterprisedevelopers #datascience #machinelearning
-
+15
-
"Orca Surpasses ChatGPT: A New Student Outshines Its Mentor" 🌟🤖 🌟 Orca, by Microsoft Research, a new chatbot showing strong performance! 🤖 It retains 95% of ChatGPT and 85% of GPT-4 quality across datasets! 📊 Orca outperforms GPT-3.5 in several tasks and is on par with TD-003. It shows a 10-point improvement over Vicuna and performs marginally better than ChatGPT. 🚀 However, GPT-4 surpasses all other models in performance. On the AGIEval framework, Orca retains 98% of ChatGPT and 89% of GPT-4 quality, outperforming Vicuna and slightly outperforming ChatGPT. But it lags GPT-4 by 26%. 🌐 Training: Trained on FLAN-5M and FLAN-1M using 20 NVIDIA A100 GPUs with 80GB memory, it took 160 hours for 4 epochs on FLAN-5M and 40 hours on FLAN-1M. Data collection from GPT-3.5-turbo (ChatGPT) and GPT-4 took 2 and 3 weeks, respectively. Excited to see how Orca keeps evolving! 🎉 #chatbots #AI #Orca #GPT4 #NVIDIA #microsoft #openai #bard #llms #azureopenai #cazton Cazton Microsoft