- 71
- 4 608
Travis Frisinger
United States
Приєднався 3 лют 2009
I'm just a guy on a journey to weave AI into my day-to-day. Ride along as I navigate the blend of triumphs and oops moments, all in pursuit of how AI can make us Better, Faster, Stronger.
Welcome to my AI Adventures.
Welcome to my AI Adventures.
AI Unplugged 0009: Vector DBs and Embeddings – The Foundation of Modern AI
In today's episode, I am diving deep into two core technologies that are shaping the future of AI: Vector Databases and Embeddings. These concepts power everything from personalized search results to recommendation systems and are essential tools for AI engineers.
I will cover:
+ How embeddings represent data in high-dimensional space, enabling AI to understand context and similarity.
+ The difference between vector search and traditional lexical search systems like Elastic search.
+ The role of graph databases like Neo4j in AI systems, and how they complement vector databases.
+ How AI engineers can leverage these technologies to build scalable, intelligent AI systems.
Understanding these technologies is crucial when you're building search engines, recommendation systems, or advanced AI applications.
HOST
@GptWithMeNow (Travis Frisinger)
HIGHLIGHTS
00:00 - Into
01:07 - What Are Embeddings
07:30 - From Lexical Search to Vector Search-A Paradigm Shift
15:13 - Embedding Models vs. Large Language Models
18:30 - Vector DBs -Storing and Retrieving Embeddings Efficiently
21:23 - Integrating Graph DBs with Vector DBs in AI Systems
24:40 - AI Applications of Vector DBs and Graph DBs
27:00 - The Takeaway
LINKS
Pinecone DB: www.pinecone.io/
Weaviate DB: weaviate.io/
FAISS DB: github.com/facebookresearch/faiss
Elasticsearch: www.elastic.co/
Neo4j: neo4j.com/
BERT-Sentence: huggingface.co/sentence-transformers
COMMUNITY
Blog: aibuddy.software/
GPT With Me: ua-cam.com/play/PL0X82GOpevvawB9w4T-WlAac-XoGBYIo_.html
Postcast: ua-cam.com/play/PL0X82GOpevvbeirz96HH4WWXmj0sJqXUj.html
I will cover:
+ How embeddings represent data in high-dimensional space, enabling AI to understand context and similarity.
+ The difference between vector search and traditional lexical search systems like Elastic search.
+ The role of graph databases like Neo4j in AI systems, and how they complement vector databases.
+ How AI engineers can leverage these technologies to build scalable, intelligent AI systems.
Understanding these technologies is crucial when you're building search engines, recommendation systems, or advanced AI applications.
HOST
@GptWithMeNow (Travis Frisinger)
HIGHLIGHTS
00:00 - Into
01:07 - What Are Embeddings
07:30 - From Lexical Search to Vector Search-A Paradigm Shift
15:13 - Embedding Models vs. Large Language Models
18:30 - Vector DBs -Storing and Retrieving Embeddings Efficiently
21:23 - Integrating Graph DBs with Vector DBs in AI Systems
24:40 - AI Applications of Vector DBs and Graph DBs
27:00 - The Takeaway
LINKS
Pinecone DB: www.pinecone.io/
Weaviate DB: weaviate.io/
FAISS DB: github.com/facebookresearch/faiss
Elasticsearch: www.elastic.co/
Neo4j: neo4j.com/
BERT-Sentence: huggingface.co/sentence-transformers
COMMUNITY
Blog: aibuddy.software/
GPT With Me: ua-cam.com/play/PL0X82GOpevvawB9w4T-WlAac-XoGBYIo_.html
Postcast: ua-cam.com/play/PL0X82GOpevvbeirz96HH4WWXmj0sJqXUj.html
Переглядів: 55
Відео
AI Unplugged 0008 : Conversational Interfaces - Chat, Tools, and Agents
Переглядів 73Місяць тому
In this episode of AI Unplugged, host Travis Frisinger continues the series, "The Rise of the AI Engineer," with a deep dive into Conversational Interfaces. From simple chatbots to advanced AI agents, conversational systems are transforming how we interact with technology. Travis covers: - The evolution of conversational interfaces from basic chat to tool-integrated systems. - How prompts are t...
AI Unplugged 0007 : The Rise of the AI Engineer - Why It’s Time to Build AI Products
Переглядів 33Місяць тому
In this episode of AI Unplugged, host Travis Frisinger kicks off an 8-part series on The Rise of the AI Engineer. Learn what sets AI engineers apart from traditional software engineers and machine learning specialists. Travis explores how pre-trained models and APIs are transforming AI product development, highlighting real-world examples like Notion, Repl.it, and Jasper. Discover why AI engine...
AI Unplugged Ep 0006 : Conversational Coding: AI's New Frontier in Software Development
Переглядів 162 місяці тому
In this episode of AI Unplugged, host Travis Freisinger explores the emerging trend of conversational coding, a new way of interacting with code that's reshaping the developer experience. Travis discusses how tools like Copilot, Codium, and Cursor are pushing the boundaries of innovation, enabling developers to focus on higher-level problem-solving and creative thinking rather than getting bogg...
AI Unplugged Ep 0005 : Vibes, TDD, and AI - Bridging Intuition and Engineering
Переглядів 82 місяці тому
In this episode of AI Unplugged, host Travis Frisinger dives into the world of 'vibes-based evaluations'-an intuitive approach to assessing AI outputs that parallels key practices in Test-Driven Development (TDD). Discover how this method can serve as a powerful tool for iterative improvement, guiding the development of robust and resilient AI systems. We'll also explore how to move beyond gut ...
AI Unplugged - Ep 0004 : AI Tools Unleashed: Google, NVIDIA, and the Next Gen of Innovation
Переглядів 82 місяці тому
In this episode of AI Unplugged, host Travis Frisinger takes you inside the AI Boulder Builders meetup, where industry giants like Google and NVIDIA, alongside innovative newcomers like Plotzy, ShowStop, and CodeYam, showcased the next generation of AI tools. From Google’s Data Science Agent, which streamlines data workflows by generating interactive Colab notebooks, to NVIDIA’s serverless infe...
AI Unplugged Ep 0003 : Creating a Metal Album with Generative AI: Mr. Fluffle's Tiny Reign of Terror
Переглядів 112 місяці тому
AI Unplugged Ep 0003 : Creating a Metal Album with Generative AI: Mr. Fluffle's Tiny Reign of Terror
The Lifecycle of Code : An Endless Journey
Переглядів 442 місяці тому
The Lifecycle of Code : An Endless Journey
AI Unplugged - Ep 0002 : Co-Creating with AI
Переглядів 142 місяці тому
AI Unplugged - Ep 0002 : Co-Creating with AI
AI Unplugged - Ep 0001 : Laying Down the Concept
Переглядів 413 місяці тому
AI Unplugged - Ep 0001 : Laying Down the Concept
Thank you for making this. I'm not sure who your intended audience is, monetization strategy, etc, but I'm finding it helpful and interesting as an experienced backend engineer bootstrapping myself into a new space. A few comments/suggestions: * You say repeatedly in your video that you've built similar systems for years taking a few moments to do something like "Before I had to solve <problem A> in this way with these constraints/tradeoffs; now I can solve it this other way instead with these benefits". E.g. anonymized case study. * When you say X (postgres) is better than Y (mysql), it would be very helpful to follow up with 1-2 sentences on why. E.g. "mysql makes fundamental design decision <bleh>, which limits its usefulness in <use-case list>". Doesn't have to be much but I'm listening for tidbits like this based on your experience. * The use-case combining a graph DB and vector DB to visualize bottlenecks in a logistics network was very interesting and helped me to better understand how the systems could be combined, same w/ the callout that semantic and key-word search still have separate spaces (unstructured vs. structured data). These points in your video are the critical moments in your videos and keep me watching. IMO, your value-add is your practical experience and knowledge. My 2 cents are to lean into that very deliberately.
Found your channel randomly. Would love to get more nitty-gritty in the details about real world solutions from a devops perspective, case studies, war stories.
An actually helpful discussion from someone that sounds like they actually build real solutions rather than drag/drop toys. What a breath of fresh air from all the low effort videos.
I am stoked that you are enjoying the discussion. I am trying document my experiences and thoughts on what is is like to work in this space. I am planning on doing some deep drive stuff after I lay out my POV on AI engineering. I want to dig a bit more on the things you pointed out like devops and war stories.
well, we gonna fix voice quality!
Yep, I tried setting some filters to remove background noise and it failed to work as expected.
Really great - love the song!
Thanks!
Great! One of our cat´s steels the socks, as she was young!😂😉
I went back after recording to compare Groq vs OpenAI and found that Groq truncated the transcript for longer recordings. It was fun but not production ready. Going to stick with OpenAI for now, but I can swap at anytime thanks to my Factory method for picking transcription providers.
I can't even express how happy I am that this exists 🤘🔥🤘
Thanks!
Funny! Nice job!
I came for the cat and stayed for the metal. I can even forgive that it's AI because cat.
King Kitty so cute. I love cats.
Nice
I read as fecal matter.
I think the vocals could stand to be balanced a little louder and having a transcript along side it would be a nice quality of life thing. But over all I really liked the song and even if the image is AI generated its pretty good.
My aunt's cat.
Very good. Looks like my neighbour's cat.
What a phenomenal tune!
Clicked because of the thumbnail, stayed because of the song 🎵
Wouldve hit harder if there was a ‘meow’ at 0:10🗣️
I'm just like a cat. I'm a goddamn survivor!
Be sure to land on your feet!
An incredible discovery! Melodic death metal, a cat and devastation? Everything I love! A great song that will be played a thousand times in my car! Thank you for this masterpiece
Thank you! I'm thrilled you enjoyed it. Rock on and keep playing it loud!
Lovely
hey, cool
🙂 "Promo SM"