Michael Nemtsev
@michaelnemtsev.bsky.social
APAC Senior Director #EPAM | Ex #Microsoft | Helping customers to become data-driven companies embracing #AI and Cloud. Opinions are mine
🚀 Avoid common AI engineering pitfalls! 🤖 From data issues to infra challenges. A detailed breakdown of the key mistakes teams make.
Don't let these slow you down! Read here: huyenchip.com/2025/01/16/a...
#AI #MachineLearning #GenAI
Don't let these slow you down! Read here: huyenchip.com/2025/01/16/a...
#AI #MachineLearning #GenAI
https://huyenchip.com/2025/01/16/ai-engineering-pitfalls.html]
March 5, 2025 at 3:17 AM
🚀 Avoid common AI engineering pitfalls! 🤖 From data issues to infra challenges. A detailed breakdown of the key mistakes teams make.
Don't let these slow you down! Read here: huyenchip.com/2025/01/16/a...
#AI #MachineLearning #GenAI
Don't let these slow you down! Read here: huyenchip.com/2025/01/16/a...
#AI #MachineLearning #GenAI
🚀 Excited to discover the Atomic AI Agent framework github.com/BrainBlend-A... 🙌 A game-changer for AI devs prioritising simplicity, modularity & control, without the LangChain overload. I didn't try it yet, but it's clean approach is super promising to be used in the next project
#AI #Agents #LLM
#AI #Agents #LLM
GitHub - BrainBlend-AI/atomic-agents: Building AI agents, atomically
Building AI agents, atomically. Contribute to BrainBlend-AI/atomic-agents development by creating an account on GitHub.
github.com
January 13, 2025 at 1:58 PM
🚀 Excited to discover the Atomic AI Agent framework github.com/BrainBlend-A... 🙌 A game-changer for AI devs prioritising simplicity, modularity & control, without the LangChain overload. I didn't try it yet, but it's clean approach is super promising to be used in the next project
#AI #Agents #LLM
#AI #Agents #LLM
🔍 "LLMs & vector embeddings are the missing piece of the NLP puzzle—not the whole picture". Relying solely on RAG and embeddings can fail.
What are your approaches to set the right context?
🔗 Read more: blog.cubed.run/the-insanity...
#AI #RAG #NLP #MachineLearning #LLM
What are your approaches to set the right context?
🔗 Read more: blog.cubed.run/the-insanity...
#AI #RAG #NLP #MachineLearning #LLM
The Insanity of Relying on Vector Embeddings: Why RAG Fails
In RAG, the goal is to locate the stored information that has the highest percentage of sameness to the provided query. Vector similarity…
blog.cubed.run
December 17, 2024 at 11:49 AM
🔍 "LLMs & vector embeddings are the missing piece of the NLP puzzle—not the whole picture". Relying solely on RAG and embeddings can fail.
What are your approaches to set the right context?
🔗 Read more: blog.cubed.run/the-insanity...
#AI #RAG #NLP #MachineLearning #LLM
What are your approaches to set the right context?
🔗 Read more: blog.cubed.run/the-insanity...
#AI #RAG #NLP #MachineLearning #LLM
Discover how to build simple yet powerful AI agents that ensure consistency & data structures using Pydantic + Llama 3. 📚 Perfect for personal bots and beyond.
🔗 Check it out: pub.towardsai.net/pydantic-ai-...
#AI #LLM #Python #Pydantic #MachineLearning
🔗 Check it out: pub.towardsai.net/pydantic-ai-...
#AI #LLM #Python #Pydantic #MachineLearning
Pydantic AI + Web Scraper + Llama 3.3 Python = Powerful AI Research Agent
In this video, I have a super quick tutorial showing you how to create a multi-agent chatbot with Pydantic AI, Web Scraper and Llama 3.3 to…
pub.towardsai.net
December 16, 2024 at 1:21 AM
Discover how to build simple yet powerful AI agents that ensure consistency & data structures using Pydantic + Llama 3. 📚 Perfect for personal bots and beyond.
🔗 Check it out: pub.towardsai.net/pydantic-ai-...
#AI #LLM #Python #Pydantic #MachineLearning
🔗 Check it out: pub.towardsai.net/pydantic-ai-...
#AI #LLM #Python #Pydantic #MachineLearning
A good example of how Agentic systems evolved over time and complexity you need to deal with
cobusgreyling.medium.com/the-evolutio...
#AI #LLM #Agentinc
cobusgreyling.medium.com/the-evolutio...
#AI #LLM #Agentinc
December 16, 2024 at 12:59 AM
A good example of how Agentic systems evolved over time and complexity you need to deal with
cobusgreyling.medium.com/the-evolutio...
#AI #LLM #Agentinc
cobusgreyling.medium.com/the-evolutio...
#AI #LLM #Agentinc
Interesting insights from @BCG on #AI adoption in the #Energy sector.
With rising energy demand driven by data centers, integrating AI across the value chain is key to boosting quality, productivity, and cutting costs ⚡
🔗 Read more: www.bcg.com/publications...
#Innovation #Sustainability #Oil&Gas
With rising energy demand driven by data centers, integrating AI across the value chain is key to boosting quality, productivity, and cutting costs ⚡
🔗 Read more: www.bcg.com/publications...
#Innovation #Sustainability #Oil&Gas
AI Adoption in Energy Should Focus on Agility, Not Algorithms
With AI and GenAI applications poised to revolutionize the oil and gas industry, companies must develop the capabilities to quickly and flexibly adopt these new technologies.
www.bcg.com
December 3, 2024 at 2:19 AM
Interesting insights from @BCG on #AI adoption in the #Energy sector.
With rising energy demand driven by data centers, integrating AI across the value chain is key to boosting quality, productivity, and cutting costs ⚡
🔗 Read more: www.bcg.com/publications...
#Innovation #Sustainability #Oil&Gas
With rising energy demand driven by data centers, integrating AI across the value chain is key to boosting quality, productivity, and cutting costs ⚡
🔗 Read more: www.bcg.com/publications...
#Innovation #Sustainability #Oil&Gas
Building a RAG with PDF sources and considering the best format structure for chunks?
🗂️ Unlock the power of Markdown and explore the top libraries for seamless PDF-to-Markdown conversion!
towardsdatascience.com/improved-rag...
#AI #LLM #RAG #Data #Markdown
🗂️ Unlock the power of Markdown and explore the top libraries for seamless PDF-to-Markdown conversion!
towardsdatascience.com/improved-rag...
#AI #LLM #RAG #Data #Markdown
Improved RAG Document Processing With Markdown
How to read and convert PDFs to Markdown for better RAG results with LLMs
towardsdatascience.com
November 28, 2024 at 3:04 AM
🚀 Building another #RAG System? Explore #LightRAG Perfect for lightweight #GenerativeAI solutions! 🌟
✅ Faster performance
✅ Simplified #AI architecture
✅ Scalable retrieval-augmented generation
✅ Open-source & easy to adapt
Details 👉 github.com/HKUDS/LightRAG
Summary: medium.com/@samarrana40...
GitHub - HKUDS/LightRAG: "LightRAG: Simple and Fast Retrieval-Augmented Generation"
"LightRAG: Simple and Fast Retrieval-Augmented Generation" - HKUDS/LightRAG
github.com
November 25, 2024 at 6:50 AM
🚀 Building another #RAG System? Explore #LightRAG Perfect for lightweight #GenerativeAI solutions! 🌟
✅ Faster performance
✅ Simplified #AI architecture
✅ Scalable retrieval-augmented generation
✅ Open-source & easy to adapt
Details 👉 github.com/HKUDS/LightRAG
Summary: medium.com/@samarrana40...
Did you know that you can actually search on your previous chat histories in ChatGPT4?
This is probably the most overlooked feature in November 2024 update. Just click on the search button and it shows the search across your entire history 💪
medium.com/@jordan_gibb...
#AI #LLM #ChatGPT #OpenAI
This is probably the most overlooked feature in November 2024 update. Just click on the search button and it shows the search across your entire history 💪
medium.com/@jordan_gibb...
#AI #LLM #ChatGPT #OpenAI
November 25, 2024 at 6:27 AM
Did you know that you can actually search on your previous chat histories in ChatGPT4?
This is probably the most overlooked feature in November 2024 update. Just click on the search button and it shows the search across your entire history 💪
medium.com/@jordan_gibb...
#AI #LLM #ChatGPT #OpenAI
This is probably the most overlooked feature in November 2024 update. Just click on the search button and it shows the search across your entire history 💪
medium.com/@jordan_gibb...
#AI #LLM #ChatGPT #OpenAI
A great article that explains types of the RAGs you are building to access complexity & data needs properly.
Explicit Facts -> Implicit Facts -> Understand Reasoning Behind Facts -> Context Based Reasoning
Source: cobusgreyling.medium.com/four-levels-...
#AI #RAG #DataPlatform #Data #LLM
Explicit Facts -> Implicit Facts -> Understand Reasoning Behind Facts -> Context Based Reasoning
Source: cobusgreyling.medium.com/four-levels-...
#AI #RAG #DataPlatform #Data #LLM
November 22, 2024 at 4:27 AM
A great article that explains types of the RAGs you are building to access complexity & data needs properly.
Explicit Facts -> Implicit Facts -> Understand Reasoning Behind Facts -> Context Based Reasoning
Source: cobusgreyling.medium.com/four-levels-...
#AI #RAG #DataPlatform #Data #LLM
Explicit Facts -> Implicit Facts -> Understand Reasoning Behind Facts -> Context Based Reasoning
Source: cobusgreyling.medium.com/four-levels-...
#AI #RAG #DataPlatform #Data #LLM
If you are building a custom RAG solution the first time, I strongly recommend to read this pub.towardsai.net/dear-it-depa...
Very detailed and hands-on experience on the hidden cost of custom RAGs and its complexity
#AI #RAG #ML #Data #DataPlatform
Very detailed and hands-on experience on the hidden cost of custom RAGs and its complexity
#AI #RAG #ML #Data #DataPlatform
Dear IT Departments, Please Stop Trying To Build Your Own RAG
IT departments convince themselves that building their own RAG-based chat is easy. It’s not. It’s a nightmare.
pub.towardsai.net
November 22, 2024 at 3:33 AM
If you are building a custom RAG solution the first time, I strongly recommend to read this pub.towardsai.net/dear-it-depa...
Very detailed and hands-on experience on the hidden cost of custom RAGs and its complexity
#AI #RAG #ML #Data #DataPlatform
Very detailed and hands-on experience on the hidden cost of custom RAGs and its complexity
#AI #RAG #ML #Data #DataPlatform
🚀 Exciting news from #MicrosoftIgnite 2024! The new Azure AI Foundry and SDK is here featuring an enhanced Azure AI Studio.
The most interesting feature is a seamless model response evaluation.
Explore the details: techcommunity.microsoft.com/blog/AIPlatf...
#AI #AzureAI #Microsoft #ML
The most interesting feature is a seamless model response evaluation.
Explore the details: techcommunity.microsoft.com/blog/AIPlatf...
#AI #AzureAI #Microsoft #ML
Ignite 2024: Announcing the Azure AI Foundry SDK | Microsoft Community Hub
The Azure AI Foundry SDK provides everything needed to build modern, intelligent applications from prototype to production.
techcommunity.microsoft.com
November 21, 2024 at 3:05 PM
🚀 Exciting news from #MicrosoftIgnite 2024! The new Azure AI Foundry and SDK is here featuring an enhanced Azure AI Studio.
The most interesting feature is a seamless model response evaluation.
Explore the details: techcommunity.microsoft.com/blog/AIPlatf...
#AI #AzureAI #Microsoft #ML
The most interesting feature is a seamless model response evaluation.
Explore the details: techcommunity.microsoft.com/blog/AIPlatf...
#AI #AzureAI #Microsoft #ML
Feeling lost in the sea of new features from Microsoft Ignite 2024? 🌊
Don't worry! Catch up with the Ignite Book of News here: news.microsoft.com/ignite-2024-... 📖✨
#MicrosoftIgnite #TechUpdates #Microsoft #Ignite2024
Don't worry! Catch up with the Ignite Book of News here: news.microsoft.com/ignite-2024-... 📖✨
#MicrosoftIgnite #TechUpdates #Microsoft #Ignite2024
Microsoft Ignite 2024 Book of News
The Book of News is your guide to all key items being announced at Microsoft Ignite 2024.
news.microsoft.com
November 20, 2024 at 3:34 AM
Feeling lost in the sea of new features from Microsoft Ignite 2024? 🌊
Don't worry! Catch up with the Ignite Book of News here: news.microsoft.com/ignite-2024-... 📖✨
#MicrosoftIgnite #TechUpdates #Microsoft #Ignite2024
Don't worry! Catch up with the Ignite Book of News here: news.microsoft.com/ignite-2024-... 📖✨
#MicrosoftIgnite #TechUpdates #Microsoft #Ignite2024
🚀 Excited to see GitHub Copilot's Code Review feature got live recently! 🤖
It brings AI-driven, agentic approaches to streamline code reviews in the DevOps process, boosting productivity and quality.
Check it out: docs.github.com/en/copilot/u...
#DevOps #AI #GitHubCopilot #CodeReview
Using GitHub Copilot code review - GitHub Docs
Learn how to request a code review from GitHub Copilot.
docs.github.com
November 19, 2024 at 7:17 AM
🚀 Excited to see GitHub Copilot's Code Review feature got live recently! 🤖
It brings AI-driven, agentic approaches to streamline code reviews in the DevOps process, boosting productivity and quality.
Check it out: docs.github.com/en/copilot/u...
#DevOps #AI #GitHubCopilot #CodeReview
ML is now easier for Go devs! 🚀 Use Python ML models (TensorFlow, JAX, PyTorch) as sidecars for seamless integration & performance.
Details here: eli.thegreenplace.net/2024/ml-in-g...
#GoLang #MachineLearning #Python #AI #DevTips
Details here: eli.thegreenplace.net/2024/ml-in-g...
#GoLang #MachineLearning #Python #AI #DevTips
ML in Go with a Python sidecar - Eli Bendersky's website
eli.thegreenplace.net
November 19, 2024 at 3:19 AM
ML is now easier for Go devs! 🚀 Use Python ML models (TensorFlow, JAX, PyTorch) as sidecars for seamless integration & performance.
Details here: eli.thegreenplace.net/2024/ml-in-g...
#GoLang #MachineLearning #Python #AI #DevTips
Details here: eli.thegreenplace.net/2024/ml-in-g...
#GoLang #MachineLearning #Python #AI #DevTips
Let's all change our approach from "data as an application" to "data as a product" when managing data to enhance quality, decision making and time-to-market
towardsdatascience.com/deliver-your...
#DataMesh #DataGovernance #Architecture
towardsdatascience.com/deliver-your...
#DataMesh #DataGovernance #Architecture
Deliver Your Data as a Product, But Not as an Application
Data as a product is an intriguing concept, but beware of the application trap
towardsdatascience.com
November 15, 2024 at 2:16 AM
Let's all change our approach from "data as an application" to "data as a product" when managing data to enhance quality, decision making and time-to-market
towardsdatascience.com/deliver-your...
#DataMesh #DataGovernance #Architecture
towardsdatascience.com/deliver-your...
#DataMesh #DataGovernance #Architecture
Did you know that there is 🚀"Azure Multimodal AI & LLM Processing Accelerator" your all-in-one solution for AI-powered tasks like content summarization, data extraction, classification, and more!
Supports text, docs, audio, images, video
techcommunity.microsoft.com/blog/azure-a...
#AI #Azure #LLM
Supports text, docs, audio, images, video
techcommunity.microsoft.com/blog/azure-a...
#AI #Azure #LLM
Introducing the Azure Multimodal AI & LLM Processing Accelerator
The Azure Multimodal AI & LLM Processing Solution Accelerator enables developers to build data processing pipelines with generative AI (GenAI) models, using...
techcommunity.microsoft.com
November 14, 2024 at 11:53 AM
Did you know that there is 🚀"Azure Multimodal AI & LLM Processing Accelerator" your all-in-one solution for AI-powered tasks like content summarization, data extraction, classification, and more!
Supports text, docs, audio, images, video
techcommunity.microsoft.com/blog/azure-a...
#AI #Azure #LLM
Supports text, docs, audio, images, video
techcommunity.microsoft.com/blog/azure-a...
#AI #Azure #LLM
New to Generative AI and Transformer Models? 🚀 Dive into this visual guide for a quick, easy-to-grasp breakdown of how Transformers work and their key concepts! 👇 poloclub.github.io/transformer-...
#AI #MachineLearning #Transformers #LLM #GenAI
#AI #MachineLearning #Transformers #LLM #GenAI
Transformer Explainer: LLM Transformer Model Visually Explained
An interactive visualization tool showing you how transformer models work in large language models (LLM) like GPT.
poloclub.github.io
November 13, 2024 at 3:12 PM
New to Generative AI and Transformer Models? 🚀 Dive into this visual guide for a quick, easy-to-grasp breakdown of how Transformers work and their key concepts! 👇 poloclub.github.io/transformer-...
#AI #MachineLearning #Transformers #LLM #GenAI
#AI #MachineLearning #Transformers #LLM #GenAI
[2/2] What stands out is "Qwen2.5.1 Coder 7B" 15GB footprint, delivering performance almost identical to LLama 3.1 405B’s 200GB footprint.
A game-changer for production systems with local models!
aider.chat/docs/leaderb...
#AI #CodeGeneration
A game-changer for production systems with local models!
aider.chat/docs/leaderb...
#AI #CodeGeneration
Aider LLM Leaderboards
Quantitative benchmarks of LLM code editing skill.
aider.chat
November 11, 2024 at 1:48 AM
[2/2] What stands out is "Qwen2.5.1 Coder 7B" 15GB footprint, delivering performance almost identical to LLama 3.1 405B’s 200GB footprint.
A game-changer for production systems with local models!
aider.chat/docs/leaderb...
#AI #CodeGeneration
A game-changer for production systems with local models!
aider.chat/docs/leaderb...
#AI #CodeGeneration
[1/2]🚀 The “Qwen2.5.1 Coder 7B” model for code generation is here! Achieving 63.9% on the Aider leaderboard, it’s close to ChatGPT but still behind leaders like Claude 3.5 Sonet and O1...
aider.chat/docs/leaderb...
#AI
aider.chat/docs/leaderb...
#AI
Aider LLM Leaderboards
Quantitative benchmarks of LLM code editing skill.
aider.chat
November 11, 2024 at 1:47 AM
[1/2]🚀 The “Qwen2.5.1 Coder 7B” model for code generation is here! Achieving 63.9% on the Aider leaderboard, it’s close to ChatGPT but still behind leaders like Claude 3.5 Sonet and O1...
aider.chat/docs/leaderb...
#AI
aider.chat/docs/leaderb...
#AI
An interesting case of how ChatGPT drove an education stock down by 99%, as customers switched to free LLM models
www.wsj.com/tech/ai/how-...
#AI #LLM #ChatGPT
www.wsj.com/tech/ai/how-...
#AI #LLM #ChatGPT
www.wsj.com
November 10, 2024 at 12:32 PM
An interesting case of how ChatGPT drove an education stock down by 99%, as customers switched to free LLM models
www.wsj.com/tech/ai/how-...
#AI #LLM #ChatGPT
www.wsj.com/tech/ai/how-...
#AI #LLM #ChatGPT
With #Copilot Studio's update to add #Agentic approach, centralized memory management is crucial. Solutions like MemGPT/Letta and CrewAI address this need.
Now we have a new opportunity integrating MemGPT via Copilot's Action. #AI
A good course to start bit.ly/3O1vQLb
Now we have a new opportunity integrating MemGPT via Copilot's Action. #AI
A good course to start bit.ly/3O1vQLb
LLMs as Operating Systems: Agent Memory
Build systems with MemGPT agents that can autonomously manage their memory.
bit.ly
November 10, 2024 at 8:24 AM
With #Copilot Studio's update to add #Agentic approach, centralized memory management is crucial. Solutions like MemGPT/Letta and CrewAI address this need.
Now we have a new opportunity integrating MemGPT via Copilot's Action. #AI
A good course to start bit.ly/3O1vQLb
Now we have a new opportunity integrating MemGPT via Copilot's Action. #AI
A good course to start bit.ly/3O1vQLb
An effective way to classify documents using LLMs without vendor lock-in, such as Azure Document Services. While these services offer robust features and cost efficiency, LLMs are advancing rapidly, challenging traditional methods
pub.towardsai.net/extractthink...
#AI #LLM #DocumentIntelligence
pub.towardsai.net/extractthink...
#AI #LLM #DocumentIntelligence
ExtractThinker: AI Document Intelligence with LLMs
Unveil the future of document intelligence with ExtractThinker, combining ORMs and LLMs
pub.towardsai.net
November 10, 2024 at 5:21 AM
An effective way to classify documents using LLMs without vendor lock-in, such as Azure Document Services. While these services offer robust features and cost efficiency, LLMs are advancing rapidly, challenging traditional methods
pub.towardsai.net/extractthink...
#AI #LLM #DocumentIntelligence
pub.towardsai.net/extractthink...
#AI #LLM #DocumentIntelligence
Trying Bluesky today, let's see how it goes
November 6, 2024 at 6:23 AM
Trying Bluesky today, let's see how it goes