AI Engineering Podcast
AI Engineering Podcast

AI Engineering Podcast

Tobias Macey

Overview
Episodes

Details

This show is your guidebook to building scalable and maintainable AI systems. You will learn how to architect AI applications, apply AI to your work, and the considerations involved in building or customizing new models. Everything that you need to know to deliver real impact and value with machine learning and artificial intelligence.

Recent Episodes

The Role Of Synthetic Data In Building Better AI Applications
FEB 16, 2025
The Role Of Synthetic Data In Building Better AI Applications
Summary<br />In this episode of the AI Engineering Podcast Ali Golshan, co-founder and CEO of Gretel.ai, talks about the transformative role of synthetic data in AI systems. Ali explains how synthetic data can be purpose-built for AI use cases, emphasizing privacy, quality, and structural stability. He highlights the shift from traditional methods to using language models, which offer enhanced capabilities in understanding data's deep structure and generating high-quality datasets. The conversation explores the challenges and techniques of integrating synthetic data into AI systems, particularly in production environments, and concludes with insights into the future of synthetic data, including its application in various industries, the importance of privacy regulations, and the ongoing evolution of AI systems.<br /><br /><br />Announcements<br /><ul><li>Hello and welcome to the AI Engineering Podcast, your guide to the fast-moving world of building scalable and maintainable AI systems</li><li>Seamless data integration into AI applications often falls short, leading many to adopt RAG methods, which come with high costs, complexity, and limited scalability. Cognee offers a better solution with its open-source semantic memory engine that automates data ingestion and storage, creating dynamic knowledge graphs from your data. Cognee enables AI agents to understand the meaning of your data, resulting in accurate responses at a lower cost. Take full control of your data in LLM apps without unnecessary overhead. Visit <a href="https://www.aiengineeringpodcast.com/cognee" target="_blank">aiengineeringpodcast.com/cognee</a> to learn more and elevate your AI apps and agents.</li><li>Your host is Tobias Macey and today I'm interviewing Ali Golshan about the role of synthetic data in building, scaling, and improving AI systems</li></ul>Interview<br /><ul><li>Introduction</li><li>How did you get involved in machine learning?</li><li>Can you start by summarizing what you mean by synthetic data in the context of this conversation?</li><li>How have the capabilities around the generation and integration of synthetic data changed across the pre- and post-LLM timelines?</li><li>What are the motivating factors that would lead a team or organization to invest in synthetic data generation capacity?</li><li>What are the main methods used for generation of synthetic data sets?<ul><li>How does that differ across open-source and commercial offerings?</li></ul></li><li>From a surface level it seems like synthetic data generation is a straight-forward exercise that can be owned by an engineering team. What are the main "gotchas" that crop up as you move along the adoption curve?<ul><li>What are the scaling characteristics of synthetic data generation as you go from prototype to production scale?</li></ul></li><li>domains/data types that are inappropriate for synthetic use cases (e.g. scientific or educational content)</li><li>managing appropriate distribution of values in the generation process</li><li>Beyond just producing large volumes of semi-random data (structured or otherwise), what are the other processes involved in the workflow of synthetic data and its integration into the different systems that consume it?</li><li>What are the most interesting, innovative, or unexpected ways that you have seen synthetic data generation used?</li><li>What are the most interesting, unexpected, or challenging lessons that you have learned while working on synthetic data generation?</li><li>When is synthetic data the wrong choice?</li><li>What do you have planned for the future of synthetic data capabilities at Gretel?</li></ul>Contact Info<br /><ul><li><a href="https://www.linkedin.com/in/ali-golshan/" target="_blank">LinkedIn</a></li></ul>Parting Question<br /><ul><li>From your perspective, what are the biggest gaps in tooling, technology, or training for AI systems today?</li></ul>Closing Announcements<br /><ul><li>Thank you for listening! Don't forget to check out our other shows. The <a href="https://www.dataengineeringpodcast.com" target="_blank">Data Engineering Podcast</a> covers the latest on modern data management. <a href="https://www.pythonpodcast.com" target="_blank">Podcast.__init__</a> covers the Python language, its community, and the innovative ways it is being used.</li><li>Visit the <a href="https://www.aiengineeringpodcast.com" target="_blank">site</a> to subscribe to the show, sign up for the mailing list, and read the show notes.</li><li>If you've learned something or tried out a project from the show then tell us about it! Email [email protected] with your story.</li><li>To help other people find the show please leave a review on <a href="https://podcasts.apple.com/us/podcast/the-machine-learning-podcast/id1626358243" target="_blank">iTunes</a> and tell your friends and co-workers.</li></ul>Links<br /><ul><li><a href="https://gretel.ai/" target="_blank">Gretel</a></li><li><a href="https://hadoop.apache.org/" target="_blank">Hadoop</a></li><li><a href="https://en.wikipedia.org/wiki/Long_short-term_memory" target="_blank">LSTM == Long Short-Term Memory</a></li><li><a href="https://en.wikipedia.org/wiki/Generative_adversarial_network" target="_blank">GAN == Generative Adversarial Network</a></li><li><a href="https://www.microsoft.com/en-us/research/publication/textbooks-are-all-you-need/" target="_blank">Textbooks are all you need</a> MSFT paper</li><li><a href="https://www.illumina.com/" target="_blank">Illumina</a></li></ul>The intro and outro music is from <a href="https://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/Tales_Of_A_Dead_Fish/Hitmans_Lovesong/" target="_blank">Hitman's Lovesong feat. Paola Graziano</a> by <a href="http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/" target="_blank">The Freak Fandango Orchestra</a>/<a href="https://creativecommons.org/licenses/by-sa/3.0/" target="_blank">CC BY-SA 3.0</a>
play-circle
54 MIN
Optimize Your AI Applications Automatically With The TensorZero LLM Gateway
JAN 22, 2025
Optimize Your AI Applications Automatically With The TensorZero LLM Gateway
Summary<br />In this episode of the AI Engineering podcast Viraj Mehta, CTO and co-founder of TensorZero, talks about the use of LLM gateways for managing interactions between client-side applications and various AI models. He highlights the benefits of using such a gateway, including standardized communication, credential management, and potential features like request-response caching and audit logging. The conversation also explores TensorZero's architecture and functionality in optimizing AI applications by managing structured data inputs and outputs, as well as the challenges and opportunities in automating prompt generation and maintaining interaction history for optimization purposes.<br /><br /><br />Announcements<br /><ul><li>Hello and welcome to the AI Engineering Podcast, your guide to the fast-moving world of building scalable and maintainable AI systems</li><li>Seamless data integration into AI applications often falls short, leading many to adopt RAG methods, which come with high costs, complexity, and limited scalability. Cognee offers a better solution with its open-source semantic memory engine that automates data ingestion and storage, creating dynamic knowledge graphs from your data. Cognee enables AI agents to understand the meaning of your data, resulting in accurate responses at a lower cost. Take full control of your data in LLM apps without unnecessary overhead. Visit <a href="https://www.aiengineeringpodcast.com/cognee" target="_blank">aiengineeringpodcast.com/cognee</a> to learn more and elevate your AI apps and agents.&nbsp;</li><li>Your host is Tobias Macey and today I'm interviewing Viraj Mehta about the purpose of an LLM gateway and his work on TensorZero</li></ul>Interview<br /><ul><li>Introduction</li><li>How did you get involved in machine learning?</li><li>What is an LLM gateway?<ul><li>What purpose does it serve in an AI application architecture?</li></ul></li><li>What are some of the different features and capabilities that an LLM gateway might be expected to provide?</li><li>Can you describe what TensorZero is and the story behind it?<ul><li>What are the core problems that you are trying to address with Tensor0 and for whom?</li></ul></li><li>One of the core features that you are offering is management of interaction history. How does this compare to the "memory" functionality offered by e.g. LangChain, Cognee, Mem0, etc.?</li><li>How does the presence of TensorZero in an application architecture change the ways that an AI engineer might approach the logic and control flows in a chat-based or agent-oriented project?</li><li>Can you describe the workflow of building with Tensor0 and some specific examples of how it feeds back into the performance/behavior of an LLM?</li><li>What are some of the ways in which the addition of Tensor0 or another LLM gateway might have a negative effect on the design or operation of an AI application?</li><li>What are the most interesting, innovative, or unexpected ways that you have seen TensorZero used?</li><li>What are the most interesting, unexpected, or challenging lessons that you have learned while working on TensorZero?</li><li>When is TensorZero the wrong choice?</li><li>What do you have planned for the future of TensorZero?</li></ul>Contact Info<br /><ul><li><a href="https://www.linkedin.com/in/virajrmehta/" target="_blank">LinkedIn</a></li></ul>Parting Question<br /><ul><li>From your perspective, what are the biggest gaps in tooling, technology, or training for AI systems today?</li></ul>Closing Announcements<br /><ul><li>Thank you for listening! Don't forget to check out our other shows. The <a href="https://www.dataengineeringpodcast.com" target="_blank">Data Engineering Podcast</a> covers the latest on modern data management. <a href="https://www.pythonpodcast.com" target="_blank">Podcast.__init__</a> covers the Python language, its community, and the innovative ways it is being used.</li><li>Visit the <a href="https://www.aiengineeringpodcast.com" target="_blank">site</a> to subscribe to the show, sign up for the mailing list, and read the show notes.</li><li>If you've learned something or tried out a project from the show then tell us about it! Email [email protected] with your story.</li><li>To help other people find the show please leave a review on <a href="https://podcasts.apple.com/us/podcast/the-machine-learning-podcast/id1626358243" target="_blank">iTunes</a> and tell your friends and co-workers.</li></ul>Links<br /><ul><li><a href="https://www.tensorzero.com/" target="_blank">TensorZero</a></li><li><a href="https://www.tensorops.ai/post/llm-gateways-in-production-centralized-access-security-and-monitoring" target="_blank">LLM Gateway</a></li><li><a href="https://www.litellm.ai/" target="_blank">LiteLLM</a></li><li><a href="https://openai.com/" target="_blank">OpenAI</a></li><li><a href="https://cloud.google.com/vertex-ai" target="_blank">Google Vertex</a></li><li><a href="https://www.anthropic.com/" target="_blank">Anthropic</a></li><li><a href="https://en.wikipedia.org/wiki/Reinforcement_learning" target="_blank">Reinforcement Learning</a></li><li><a href="https://en.wikipedia.org/wiki/Tokamak" target="_blank">Tokamak Reactor</a></li><li><a href="https://scholar.google.com/citations?view_op=view_citation&amp;hl=en&amp;user=4pHjHBkAAAAJ&amp;sortby=pubdate&amp;citation_for_view=4pHjHBkAAAAJ:M3ejUd6NZC8C" target="_blank">Viraj RLHF Paper</a></li><li><a href="https://arxiv.org/abs/1502.06362" target="_blank">Contextual Dueling Bandits</a></li><li><a href="https://www.superannotate.com/blog/direct-preference-optimization-dpo" target="_blank">Direct Preference Optimization</a></li><li><a href="https://en.wikipedia.org/wiki/Partially_observable_Markov_decision_process" target="_blank">Partially Observable Markov Decision Process</a></li><li><a href="https://dspy.ai/" target="_blank">DSPy</a></li><li><a href="https://pytorch.org/" target="_blank">PyTorch</a></li><li><a href="https://www.cognee.ai" target="_blank">Cognee</a></li><li><a href="https://github.com/mem0ai/mem0" target="_blank">Mem0</a></li><li><a href="https://www.langchain.com/langgraph" target="_blank">LangGraph</a></li><li><a href="https://en.wikipedia.org/wiki/Douglas_Hofstadter" target="_blank">Douglas Hofstadter</a></li><li><a href="https://github.com/openai/gym" target="_blank">OpenAI Gym</a></li><li><a href="https://en.wikipedia.org/wiki/OpenAI_o1" target="_blank">OpenAI o1</a></li><li><a href="https://en.wikipedia.org/wiki/OpenAI_o3" target="_blank">OpenAI o3</a></li><li><a href="https://www.promptingguide.ai/techniques/cot" target="_blank">Chain Of Thought</a></li></ul>The intro and outro music is from <a href="https://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/Tales_Of_A_Dead_Fish/Hitmans_Lovesong/" target="_blank">Hitman's Lovesong feat. Paola Graziano</a> by <a href="http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/" target="_blank">The Freak Fandango Orchestra</a>/<a href="https://creativecommons.org/licenses/by-sa/3.0/" target="_blank">CC BY-SA 3.0</a>
play-circle
63 MIN
Harnessing The Engine Of AI
DEC 16, 2024
Harnessing The Engine Of AI
Summary<br />In this episode of the AI Engineering Podcast Ron Green, co-founder and CTO of KungFu AI, talks about the evolving landscape of AI systems and the challenges of harnessing generative AI engines. Ron shares his insights on the limitations of large language models (LLMs) as standalone solutions and emphasizes the need for human oversight, multi-agent systems, and robust data management to support AI initiatives. He discusses the potential of domain-specific AI solutions, RAG approaches, and mixture of experts to enhance AI capabilities while addressing risks. The conversation also explores the evolving AI ecosystem, including tooling and frameworks, strategic planning, and the importance of interpretability and control in AI systems. Ron expresses optimism about the future of AI, predicting significant advancements in the next 20 years and the integration of AI capabilities into everyday software applications.<br /><br /><br />Announcements<br /><ul><li>Hello and welcome to the AI Engineering Podcast, your guide to the fast-moving world of building scalable and maintainable AI systems</li><li>Seamless data integration into AI applications often falls short, leading many to adopt RAG methods, which come with high costs, complexity, and limited scalability. Cognee offers a better solution with its open-source semantic memory engine that automates data ingestion and storage, creating dynamic knowledge graphs from your data. Cognee enables AI agents to understand the meaning of your data, resulting in accurate responses at a lower cost. Take full control of your data in LLM apps without unnecessary overhead. Visit <a href="https://www.aiengineeringpodcast.com/cognee" target="_blank">aiengineeringpodcast.com/cognee</a> to learn more and elevate your AI apps and agents.&nbsp;</li><li>Your host is Tobias Macey and today I'm interviewing Ron Green about the wheels that we need for harnessing the power of the generative AI engine</li></ul>Interview<br /><ul><li>Introduction</li><li>How did you get involved in machine learning?</li><li>Can you describe what you see as the main shortcomings of LLMs as a stand-alone solution (to anything)?</li><li>The most established vehicle for harnessing LLM capabilities is the RAG pattern. What are the main limitations of that as a "product" solution?</li><li>The idea of multi-agent or mixture-of-experts systems is a more sophisticated approach that is gaining some attention. What do you see as the pro/con conversation around that pattern?</li><li>Beyond the system patterns that are being developed there is also a rapidly shifting ecosystem of frameworks, tools, and point solutions that plugin to various points of the AI lifecycle. How does that volatility hinder the adoption of generative AI in different contexts?<ul><li>In addition to the tooling, the models themselves are rapidly changing. How much does that influence the ways that organizations are thinking about whether and when to test the waters of AI?</li></ul></li><li>Continuing on the metaphor of LLMs and engines and the need for vehicles, where are we on the timeline in relation to the model T Ford?<ul><li>What are the vehicle categories that we still need to design and develop? (e.g. sedans, mini-vans, freight trucks, etc.)</li></ul></li><li>The current transformer architecture is starting to reach scaling limits that lead to diminishing returns. Given your perspective as an industry veteran, what are your thoughts on the future trajectory of AI model architectures?<ul><li>What is the ongoing role of regression style ML in the landscape of generative AI?</li></ul></li><li>What are the most interesting, innovative, or unexpected ways that you have seen LLMs used to power a "vehicle"?</li><li>What are the most interesting, unexpected, or challenging lessons that you have learned while working in this phase of AI?</li><li>When is generative AI/LLMs the wrong choice?</li></ul>Contact Info<br /><ul><li><a href="https://www.linkedin.com/in/rongreen/" target="_blank">LinkedIn</a></li></ul>Parting Question<br /><ul><li>From your perspective, what are the biggest gaps in tooling, technology, or training for AI systems today?</li></ul>Closing Announcements<br /><ul><li>Thank you for listening! Don't forget to check out our other shows. The <a href="https://www.dataengineeringpodcast.com" target="_blank">Data Engineering Podcast</a> covers the latest on modern data management. <a href="https://www.pythonpodcast.com" target="_blank">Podcast.__init__</a> covers the Python language, its community, and the innovative ways it is being used.</li><li>Visit the <a href="https://www.aiengineeringpodcast.com" target="_blank">site</a> to subscribe to the show, sign up for the mailing list, and read the show notes.</li><li>If you've learned something or tried out a project from the show then tell us about it! Email [email protected] with your story.</li><li>To help other people find the show please leave a review on <a href="https://podcasts.apple.com/us/podcast/the-machine-learning-podcast/id1626358243" target="_blank">iTunes</a> and tell your friends and co-workers.</li></ul>Links<br /><ul><li><a href="https://www.kungfu.ai/" target="_blank">Kungfu.ai</a></li><li><a href="https://www.llama.com/" target="_blank">Llama</a> open generative AI models</li><li><a href="https://openai.com/index/chatgpt/" target="_blank">ChatGPT</a></li><li><a href="https://github.com/features/copilot" target="_blank">Copilot</a></li><li><a href="https://www.cursor.com/" target="_blank">Cursor</a></li><li><a href="https://en.wikipedia.org/wiki/Retrieval-augmented_generation" target="_blank">RAG == Retrieval Augmented Generation</a><ul><li><a href="https://www.aiengineeringpodcast.com/retrieval-augmented-generation-implementation-episode-34" target="_blank">Podcast Episode</a></li></ul></li><li><a href="https://huggingface.co/blog/moe" target="_blank">Mixture of Experts</a></li><li><a href="https://en.wikipedia.org/wiki/Deep_learning" target="_blank">Deep Learning</a></li><li><a href="https://en.wikipedia.org/wiki/Random_forest" target="_blank">Random Forest</a></li><li><a href="https://en.wikipedia.org/wiki/Supervised_learning" target="_blank">Supervised Learning</a></li><li><a href="https://en.wikipedia.org/wiki/Active_learning_(machine_learning" target="_blank">Active Learning</a>)</li><li><a href="https://yann.lecun.com/" target="_blank">Yann LeCunn</a></li><li><a href="https://en.wikipedia.org/wiki/Reinforcement_learning_from_human_feedback" target="_blank">RLHF == Reinforcement Learning from Human Feedback</a></li><li><a href="https://en.wikipedia.org/wiki/Ford_Model_T" target="_blank">Model T Ford</a></li><li><a href="https://github.com/state-spaces/mamba" target="_blank">Mamba selective state space</a></li><li><a href="https://news.mit.edu/2021/machine-learning-adapts-0128" target="_blank">Liquid Network</a></li><li><a href="https://www.promptingguide.ai/techniques/cot" target="_blank">Chain of thought</a></li><li><a href="https://openai.com/o1/" target="_blank">OpenAI o1</a></li><li><a href="https://en.wikipedia.org/wiki/Marvin_Minsky" target="_blank">Marvin Minsky</a></li><li><a href="https://en.wikipedia.org/wiki/Von_Neumann_architecture" target="_blank">Von Neumann Architecture</a></li><li><a href="https://arxiv.org/abs/1706.03762" target="_blank">Attention Is All You Need</a></li><li><a href="https://en.wikipedia.org/wiki/Multilayer_perceptron" target="_blank">Multilayer Perceptron</a></li><li><a href="https://builtin.com/data-science/dot-product-matrix" target="_blank">Dot Product</a></li><li><a href="https://en.wikipedia.org/wiki/Diffusion_model" target="_blank">Diffusion Model</a></li><li><a href="https://en.wikipedia.org/wiki/Gaussian_noise" target="_blank">Gaussian Noise</a></li><li><a href="https://blog.google/technology/ai/google-deepmind-isomorphic-alphafold-3-ai-model/" target="_blank">AlphaFold 3</a></li><li><a href="https://www.anthropic.com/" target="_blank">Anthropic</a></li><li><a href="https://paperswithcode.com/method/sparse-autoencoder" target="_blank">Sparse Autoencoder</a></li></ul>The intro and outro music is from <a href="https://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/Tales_Of_A_Dead_Fish/Hitmans_Lovesong/" target="_blank">Hitman's Lovesong feat. Paola Graziano</a> by <a href="http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/" target="_blank">The Freak Fandango Orchestra</a>/<a href="https://creativecommons.org/licenses/by-sa/3.0/" target="_blank">CC BY-SA 3.0</a>
play-circle
55 MIN
The Complex World of Generative AI Governance
DEC 1, 2024
The Complex World of Generative AI Governance
Summary<br />In this episode of the AI Engineering Podcast Jim Olsen, CTO of ModelOp, talks about the governance of generative AI models and applications. Jim shares his extensive experience in software engineering and machine learning, highlighting the importance of governance in high-risk applications like healthcare. He explains that governance is more about the use cases of AI models rather than the models themselves, emphasizing the need for proper inventory and monitoring to ensure compliance and mitigate risks. The conversation covers challenges organizations face in implementing AI governance policies, the importance of technical controls for data governance, and the need for ongoing monitoring and baselines to detect issues like PII disclosure and model drift. Jim also discusses the balance between innovation and regulation, particularly with evolving regulations like those in the EU, and provides valuable perspectives on the current state of AI governance and the need for robust model lifecycle management.<br /><br /><br />Announcements<br /><ul><li>Hello and welcome to the AI Engineering Podcast, your guide to the fast-moving world of building scalable and maintainable AI systems</li><li>Your host is Tobias Macey and today I'm interviewing Jim Olsen about governance of your generative AI models and applications</li></ul>Interview<br /><ul><li>Introduction</li><li>How did you get involved in machine learning?</li><li>Can you describe what governance means in the context of generative AI models? (e.g. governing the models, their applications, their outputs, etc.)</li><li>Governance is typically a hybrid endeavor of technical and organizational policy creation and enforcement. From the organizational perspective, what are some of the difficulties that teams are facing in understanding what those policies need to encompass?<ul><li>How much familiarity with the capabilities and limitations of the models is necessary to engage productively with policy debates?</li></ul></li><li>The regulatory landscape around AI is still very nascent. Can you give an overview of the current state of legal burden related to AI?<ul><li>What are some of the regulations that you consider necessary but as-of-yet absent?</li></ul></li><li>Data governance as a practice typically relates to controls over who can access what information and how it can be used. The controls for those policies are generally available in the data warehouse, business intelligence, etc. What are the different dimensions of technical controls that are needed in the application of generative AI systems?<ul><li>How much of the controls that are present for governance of analytical systems are applicable to the generative AI arena?</li></ul></li><li>What are the elements of risk that change when considering internal vs. consumer facing applications of generative AI?<ul><li>How do the modalities of the AI models impact the types of risk that are involved? (e.g. language vs. vision vs. audio)</li></ul></li><li>What are some of the technical aspects of the AI tools ecosystem that are in greatest need of investment to ease the burden of risk and validation of model use?</li><li>What are the most interesting, innovative, or unexpected ways that you have seen AI governance implemented?</li><li>What are the most interesting, unexpected, or challenging lessons that you have learned while working on AI governance?</li><li>What are the technical, social, and organizational trends of AI risk and governance that you are monitoring?</li></ul>Contact Info<br /><ul><li><a href="https://www.linkedin.com/in/jimolsen/" target="_blank">LinkedIn</a></li></ul>Parting Question<br /><ul><li>From your perspective, what are the biggest gaps in tooling, technology, or training for AI systems today?</li></ul>Closing Announcements<br /><ul><li>Thank you for listening! Don't forget to check out our other shows. The <a href="https://www.dataengineeringpodcast.com" target="_blank">Data Engineering Podcast</a> covers the latest on modern data management. <a href="https://www.pythonpodcast.com" target="_blank">Podcast.__init__</a> covers the Python language, its community, and the innovative ways it is being used.</li><li>Visit the <a href="https://www.aiengineeringpodcast.com" target="_blank">site</a> to subscribe to the show, sign up for the mailing list, and read the show notes.</li><li>If you've learned something or tried out a project from the show then tell us about it! Email [email protected] with your story.</li><li>To help other people find the show please leave a review on <a href="https://podcasts.apple.com/us/podcast/the-machine-learning-podcast/id1626358243" target="_blank">iTunes</a> and tell your friends and co-workers.</li></ul>Links<br /><ul><li><a href="https://www.modelop.com/" target="_blank">ModelOp</a></li><li><a href="https://en.wikipedia.org/wiki/Foundation_model" target="_blank">Foundation Models</a></li><li><a href="https://en.wikipedia.org/wiki/General_Data_Protection_Regulation" target="_blank">GDPR</a></li><li><a href="https://www.europarl.europa.eu/topics/en/article/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence" target="_blank">EU AI Regulation</a></li><li><a href="https://www.llama.com/llama2/" target="_blank">Llama 2</a></li><li><a href="https://aws.amazon.com/bedrock/" target="_blank">AWS Bedrock</a></li><li><a href="https://en.wikipedia.org/wiki/Shadow_IT" target="_blank">Shadow IT</a></li><li><a href="https://en.wikipedia.org/wiki/Retrieval-augmented_generation" target="_blank">RAG == Retrieval Augmented Generation</a><ul><li><a href="https://www.aiengineeringpodcast.com/retrieval-augmented-generation-implementation-episode-34" target="_blank">Podcast Episode</a></li></ul></li><li><a href="https://github.com/NVIDIA/NeMo" target="_blank">Nvidia NEMO</a></li><li><a href="https://www.langchain.com/" target="_blank">LangChain</a></li><li><a href="https://shap.readthedocs.io/en/latest/example_notebooks/overviews/An%20introduction%20to%20explainable%20AI%20with%20Shapley%20values.html" target="_blank">Shapley Values</a></li><li><a href="https://llm-guard.com/output_scanners/gibberish/" target="_blank">Gibberish Detection</a></li></ul>The intro and outro music is from <a href="https://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/Tales_Of_A_Dead_Fish/Hitmans_Lovesong/" target="_blank">Hitman's Lovesong feat. Paola Graziano</a> by <a href="http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/" target="_blank">The Freak Fandango Orchestra</a>/<a href="https://creativecommons.org/licenses/by-sa/3.0/" target="_blank">CC BY-SA 3.0</a>
play-circle
54 MIN
Building Semantic Memory for AI With Cognee
NOV 25, 2024
Building Semantic Memory for AI With Cognee
Summary<br />In this episode of the AI Engineering Podcast, Vasilije Markovich talks about enhancing Large Language Models (LLMs) with memory to improve their accuracy. He discusses the concept of memory in LLMs, which involves managing context windows to enhance reasoning without the high costs of traditional training methods. He explains the challenges of forgetting in LLMs due to context window limitations and introduces the idea of hierarchical memory, where immediate retrieval and long-term information storage are balanced to improve application performance. Vasilije also shares his work on Cognee, a tool he's developing to manage semantic memory in AI systems, and discusses its potential applications beyond its core use case. He emphasizes the importance of combining cognitive science principles with data engineering to push the boundaries of AI capabilities and shares his vision for the future of AI systems, highlighting the role of personalization and the ongoing development of Cognee to support evolving AI architectures.<br /><br /><br />Announcements<br /><ul><li>Hello and welcome to the AI Engineering Podcast, your guide to the fast-moving world of building scalable and maintainable AI systems</li><li>Your host is Tobias Macey and today I'm interviewing Vasilije Markovic about adding memory to LLMs to improve their accuracy</li></ul>Interview<br /><ul><li>Introduction</li><li>How did you get involved in machine learning?</li><li>Can you describe what "memory" is in the context of LLM systems?</li><li>What are the symptoms of "forgetting" that manifest when interacting with LLMs?<ul><li>How do these issues manifest between single-turn vs. multi-turn interactions?</li></ul></li><li>How does the lack of hierarchical and evolving memory limit the capabilities of LLM systems?</li><li>What are the technical/architectural requirements to add memory to an LLM system/application?</li><li>How does Cognee help to address the shortcomings of current LLM/RAG architectures?</li><li>Can you describe how Cognee is implemented?<ul><li>Recognizing that it has only existed for a short time, how have the design and scope of Cognee evolved since you first started working on it?</li></ul></li><li>What are the data structures that are most useful for managing the memory structures?</li><li>For someone who wants to incorporate Cognee into their LLM architecture, what is involved in integrating it into their applications?<ul><li>How does it change the way that you think about the overall requirements for an LLM application?</li></ul></li><li>For systems that interact with multiple LLMs, how does Cognee manage context across those systems? (e.g. different agents for different use cases)</li><li>There are other systems that are being built to manage user personalization in LLm applications, how do the goals of Cognee relate to those use cases? (e.g. Mem0 - <a href="https://github.com/mem0ai/mem0)" target="_blank">https://github.com/mem0ai/mem0)</a></li><li>What are the unknowns that you are still navigating with Cognee?</li><li>What are the most interesting, innovative, or unexpected ways that you have seen Cognee used?</li><li>What are the most interesting, unexpected, or challenging lessons that you have learned while working on Cognee?</li><li>When is Cognee the wrong choice?</li><li>What do you have planned for the future of Cognee?</li></ul>Contact Info<br /><ul><li><a href="https://www.linkedin.com/in/vasilije-markovic-13302471/" target="_blank">LinkedIn</a></li></ul>Parting Question<br /><ul><li>From your perspective, what are the biggest gaps in tooling, technology, or training for AI systems today?</li></ul>Closing Announcements<br /><ul><li>Thank you for listening! Don't forget to check out our other shows. The <a href="https://www.dataengineeringpodcast.com" target="_blank">Data Engineering Podcast</a> covers the latest on modern data management. <a href="https://www.pythonpodcast.com" target="_blank">Podcast.__init__</a> covers the Python language, its community, and the innovative ways it is being used.</li><li>Visit the <a href="https://www.aiengineeringpodcast.com" target="_blank">site</a> to subscribe to the show, sign up for the mailing list, and read the show notes.</li><li>If you've learned something or tried out a project from the show then tell us about it! Email [email protected] with your story.</li><li>To help other people find the show please leave a review on <a href="https://podcasts.apple.com/us/podcast/the-machine-learning-podcast/id1626358243" target="_blank">iTunes</a> and tell your friends and co-workers.</li></ul>Links<br /><ul><li><a href="https://www.cognee.ai/" target="_blank">Cognee</a></li><li><a href="https://en.wikipedia.org/wiki/Montenegro" target="_blank">Montenegro</a></li><li><a href="https://en.wikipedia.org/wiki/Catastrophic_interference" target="_blank">Catastrophic Forgetting</a></li><li><a href="https://poly.ai/blog/multi-turn-conversations-what-are-they-and-why-do-they-matter-for-your-customers/" target="_blank">Multi-Turn Interaction</a></li><li><a href="https://en.wikipedia.org/wiki/Retrieval-augmented_generation" target="_blank">RAG == Retrieval Augmented Generation</a><ul><li><a href="https://www.aiengineeringpodcast.com/retrieval-augmented-generation-implementation-episode-34" target="_blank">Podcast Episode</a></li></ul></li><li><a href="https://neo4j.com/blog/graphrag-manifesto/" target="_blank">GraphRAG</a><ul><li><a href="https://www.aiengineeringpodcast.com/graphrag-knowledge-graph-semantic-retrieval-episode-37" target="_blank">Podcast Episode</a></li></ul></li><li><a href="https://en.wikipedia.org/wiki/Long-term_memory" target="_blank">Long-term memory</a></li><li><a href="https://en.wikipedia.org/wiki/Short-term_memory" target="_blank">Short-term memory</a></li><li><a href="https://www.langchain.com/" target="_blank">Langchain</a></li><li><a href="https://www.llamaindex.ai/" target="_blank">LlamaIndex</a></li><li><a href="https://haystack.deepset.ai/" target="_blank">Haystack</a></li><li><a href="https://dlthub.com/" target="_blank">dlt</a><ul><li><a href="https://www.dataengineeringpodcast.com/dlt-pure-python-data-integration-episode-441" target="_blank">Data Engineering Podcast Episode</a></li></ul></li><li><a href="https://www.pinecone.io/" target="_blank">Pinecone</a><ul><li><a href="https://www.dataengineeringpodcast.com/pinecone-vector-database-similarity-search-episode-189/" target="_blank">Podcast Episode</a></li></ul></li><li><a href="https://weaviate.io/blog/what-is-agentic-rag" target="_blank">Agentic RAG</a></li><li><a href="https://airflow.apache.org/" target="_blank">Airflow</a></li><li><a href="https://en.wikipedia.org/wiki/Directed_acyclic_graph" target="_blank">DAG == Directed Acyclic Graph</a></li><li><a href="https://github.com/FalkorDB/falkordb" target="_blank">FalkorDB</a></li><li><a href="https://neo4j.com/" target="_blank">Neo4J</a></li><li><a href="https://pydantic.dev/" target="_blank">Pydantic</a></li><li><a href="https://aws.amazon.com/ecs/" target="_blank">AWS ECS</a></li><li><a href="https://aws.amazon.com/sns/" target="_blank">AWS SNS</a></li><li><a href="https://aws.amazon.com/sqs/" target="_blank">AWS SQS</a></li><li><a href="https://aws.amazon.com/lambda/" target="_blank">AWS Lambda</a></li><li><a href="https://www.evidentlyai.com/llm-guide/llm-as-a-judge" target="_blank">LLM As Judge</a></li><li><a href="https://www.mem0.ai/" target="_blank">Mem0</a></li><li><a href="https://qdrant.tech/" target="_blank">QDrant</a></li><li><a href="https://lancedb.com/" target="_blank">LanceDB</a></li><li><a href="https://duckdb.org/" target="_blank">DuckDB</a></li></ul>The intro and outro music is from <a href="https://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/Tales_Of_A_Dead_Fish/Hitmans_Lovesong/" target="_blank">Hitman's Lovesong feat. Paola Graziano</a> by <a href="http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/" target="_blank">The Freak Fandango Orchestra</a>/<a href="https://creativecommons.org/licenses/by-sa/3.0/" target="_blank">CC BY-SA 3.0</a>
play-circle
55 MIN