Agentic Conversations (formally mlops.community)

Andy Pernsteiner is the Field CTO at VAST Data, working on large-scale AI infrastructure, serverless compute near data, and the rollout of VAST’s AI Operating System.

The GPU Uptime Battle // MLOps Podcast #346 with Andy Pernsteiner, Field CTO of VAST Data.Huge thanks to VAST Data for supporting this episode!

Join the Community:

https://go.mlops.community/YTJoinIn

Get the newsletter:

https://go.mlops.community/YTNewsletter

// Abstract

Most AI projects don’t fail because of bad models; they fail because of bad data plumbing. Andy Pernsteiner joins the podcast to talk about what it actually takes to build production-grade AI systems that aren’t held together by brittle ETL scripts and data copies. He unpacks why unifying data - rather than moving it - is key to real-time, secure inference, and how event-driven, Kubernetes-native pipelines are reshaping the way developers build AI applications. It’s a conversation about cutting out the complexity, keeping data live, and building systems smart enough to keep up with your models.

// Bio

Andy is the Field Chief Technology Officer at VAST, helping customers build, deploy, and scale some of the world’s largest and most demanding computing environments.

Andy has spent the past 15 years focused on supporting and building large-scale, high-performance data platform solutions. From humble beginnings as an escalations engineer at pre-IPO Isilon, to leading a team of technical Ninjas at MapR, he’s consistently been in the frontlines solving some of the toughest challenges that customers face when implementing Big Data Analytics and next-generation AI solutions.

// Related Links

Website: www.vastdata.com

https://www.youtube.com/watch?v=HYIEgFyHaxk

https://www.youtube.com/watch?v=RyDHIMniLro

The Mom Test by Rob Fitzpatrick: https://www.momtestbook.com/

~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~

Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExplore

Join our Slack community

[https://go.mlops.community/slack]

Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)]

Sign up for the next meetup: [https://go.mlops.community/register]

MLOps Swag/Merch: [https://shop.mlops.community/]

Connect with Demetrios on LinkedIn: /dpbrinkm

Connect with Andy on LinkedIn: /andypernsteiner

Timestamps:

[00:00] Prototype to production gap

[00:21] AI expectations vs reality

[03:00] Prototype vs production costs

[07:47] Technical debt awareness

[10:13] The Mom Test

[15:40] Chaos engineering

[22:25] Data messiness reflection

[26:50] Small data value

[30:53] Platform engineer mindset shift

[34:26] Gradient description comparison

[38:12] Empathy in MLOps

[45:48] Empathy in Engineering

[51:04] GPU clusters rolling updates

[1:03:14] Checkpointing strategy comparison

[1:09:44] Predictive vs Generative AI

[1:17:51] On Growth, Community, and New Directions

[1:24:21] UX of agents

[1:32:05] Wrap up

Dr. Jeff Schwartzentruber is a Senior Machine Learning Scientist at eSentire, working on anomaly detection pipelines and the use of large language models to enhance cybersecurity operations.

The Evolution of AI in Cyber Security // MLOps Podcast #344 with Jeff Schwartzentruber, Staff Machine Learning Scientist at eSentire.

Join the Community:

https://go.mlops.community/YTJoinIn

Get the newsletter: https://go.mlops.community/YTNewsletter

// Abstract

Modern cyber operations can feel opaque. This talk explains—step by step—what a security operations center (SOC) actually does, how telemetry flows in from networks, endpoints, and cloud apps, and what an investigation can credibly reveal about attacker behavior, exposure, and control gaps. We then trace how AI has shown up in the SOC: from rules and classic machine learning for detection to natural-language tools that summarize alerts and turn questions like “show failed logins from new countries in the last 24 hours” into fast database queries. The core of the talk is our next step: agentic investigations. These GenAI agents plan their work, run queries across tools, cite evidence, and draft analyst-grade findings—with guardrails and a human in the loop. We close with what’s next: risk-aware auto-remediation, verifiable knowledge sources, and a practical checklist for adopting these capabilities safely.

// Bio

Dr. Jeff Schwartzentruber holds the position of Sr. Machine Learning Scientist at eSentire – a Canadian cybersecurity company specializing in Managed Detection and Response (MDR). Dr. Schwartzentruber’s primary academic and industry research has been concentrated on solving problems at the intersection of cybersecurity and machine learning (ML). Over his +10-year career, Dr. Schwartzentruber has been involved in applying ML for threat detection and security analytics for several large Canadian financial institutions, public sector organizations (federal), and SME’s. In addition to his private sector work, Dr. Schwartzentruber is also an Adjunct Faculty at Dalhousie University in the Department of Computer Science, a Special Graduate Faculty member with the School of Computer Science at the University of Guelph, and a Sr. Advisor on AI at the Rogers Cyber Secure Catalysts.

// Related Links

Website: https://www.esentire.com/

~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~

Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExplore

Join our Slack community [https://go.mlops.community/slack]

Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)]

Sign up for the next meetup: [https://go.mlops.community/register]

MLOps Swag/Merch: [https://shop.mlops.community/]

Connect with Demetrios on LinkedIn: /dpbrinkm

Connect with Jeff on LinkedIn: /jeff-schwartzentruber/

Jaipal Singh Goud is the CTO at Prem AI, working on model customization and privacy-preserving compute.

This episode was recorded at the Plan B studios in Lugano, Switzerland. For more information, visit https://pow.space/

How do fine-tuned models and RAG systems power personalized AI agents that learn, collaborate, and transform enterprise workflows? What kind of technical challenges do we need to first examine before this becomes real?

Demetrios Brinkmann - Founder of MLOps Community

~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~

Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExplore

Join our Slack community [https://go.mlops.community/slack]

Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)]

Sign up for the next meetup: [https://go.mlops.community/register]

MLOps Swag/Merch: [https://shop.mlops.community/]

The Semantic Layer and AI Agents // MLOps Podcast #343 with David Jayatillake, VP of AI at Cube.dev.

Join the Community:

https://go.mlops.community/YTJoinIn

Get the newsletter: https://go.mlops.community/YTNewsletter

// Abstract

David Jayatillake argues that the real battle in data isn’t about AI — it’s about who controls the semantics. In this episode, he calls out how proprietary BI tools quietly lock companies into their ecosystems, making data less open and less useful. David and Demetrios debate whether semantic layers should live in open-source hands and how AI agents might soon replace entire chunks of manual data engineering. From feature stores to LLM-driven analytics, this conversation challenges how we think about ownership, access, and the future of data workflows.

// Bio

Experienced and world-renowned data, technology, and AI leader. Expert in the application of LLMs to the semantic layer.

Writes at davidsj.substack.com about data, leadership, architecture, venture capital, and artificial intelligence.

Two-time co-founder in the data space. Founded Delphi Labs, which focused on applying LLMs to semantic layers to enable data democratization.

Regular data conference, podcast, panel, and webinar speaker.

// Related Links

Website: davidsj.substack.com

~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~

Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExplore

Join our Slack community [https://go.mlops.community/slack]

Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)]

Sign up for the next meetup: [https://go.mlops.community/register]

MLOps Swag/Merch: [https://shop.mlops.community/]

Connect with Demetrios on LinkedIn: /dpbrinkm

Connect with David on LinkedIn: /david-jayatillake/

Building Claude Code: Origin, Story, Product Iterations, & What's Next // MLOps Podcast #342 with Siddharth Bidasaria, Member of Technical Staff at Anthropic.

Join the Community:

https://go.mlops.community/YTJoinIn

Get the newsletter: https://go.mlops.community/YTNewsletter

// Abstract

Demetrios Brinkmann talks with Siddharth Bidasaria about Anthropic’s Claude code — how it was built, key features like file tools and Spotify control, and the team’s lean, user-focused approach. They explore testing, subagents, and the future of agentic coding, plus how users are pushing its limits.

// BioSoftware engineer. Founding team of Claude Code. Ex-Robinhood and Rubrik.

// Related Links

Bio: https://sidb.io/

Sid's Blog: https://sidb.io/posts/

I Let An AI Play Pokémon! - Claude plays Pokémon Creator: https://youtu.be/nRHeGJwVP18

How Data Platforms Affect ML & AI // Jake Watson // MLOps Podcast #207: https://youtu.be/xWApMuyct_4

The Agent Landscape - Lessons Learned Putting Agents Into Production: https://youtu.be/lRGldru7ohU

~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~

Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExplore

Join our Slack community [https://go.mlops.community/slack]

Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)]

Sign up for the next meetup: [https://go.mlops.community/register]

MLOps Swag/Merch: [https://shop.mlops.community/]

Connect with Demetrios on LinkedIn: /dpbrinkm

Connect with Marco on LinkedIn: /siddharthbidasaria/

Timestamps:

[00:00] MCP servers usage creativity

[00:34] Claude's code origin story

[05:17] R&D freedom and tools

[09:08] Model potential discovery

[12:06] Model adaptation strategies

[19:13] Steerability vs pattern alignment

[22:09] Features to delete

[24:12] Moore's law in LLMs

[32:42] Power user surprises

[35:56] Sub-agent evolution insights

[39:54] Agent communication governance

[45:26] At-scale agent coordination

[49:56] Wrap up

What if AI could actually remember like humans do?

Biswaroop Bhattacharjee joins Demetrios Brinkmann to challenge how we think about memory in AI. From building Cortex—a system inspired by human cognition—to exploring whether AI should forget, this conversation questions the limits of agentic memory and how far we should go in mimicking the mind.

Guest speaker: Biswaroop Bhattacharjee - Senior ML Engineer at Prem AI

Host :Demetrios Brinkmann - Founder of MLOps Community

~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~

Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExplore

Join our Slack community [https://go.mlops.community/slack]

Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)]

Sign up for the next meetup: [https://go.mlops.community/register]

MLOps Swag/Merch: [https://shop.mlops.community/]

#podcast #aiinfrastructure #aiagents #memory

LLMs at Scale: Infrastructure That Keeps AI Safe, Smart & Affordable // MLOps Podcast #341 with Marco Palladino, Kong's Co-Founder and CTO.

Join the Community: https://go.mlops.community/YTJoinIn

Get the newsletter: https://go.mlops.community/YTNewsletter

// Abstract

While conversations around AI regulations continue to evolve, the responsibility for AI continues to be with developers. In this episode, Marco Palladino, CTO and co-founder of Kong Inc., explores what it means to build and scale AI responsibly when the rulebook is still being written. He explains that infrastructure should be the frontline defense for enforcing governance, security, and reliability in AI deployments. Marco shares how Kong’s technologies, including AI Gateway and AI Manager, help organizations rein in shadow AI, reduce LLM hallucinations, improve observability, and act as the foundation for agentic workflows.

// Bio

Marco Palladino is an inventor, software developer, and internet entrepreneur. As the CTO and co-founder of Kong, he is Kong’s co-author, responsible for the design and delivery of the company’s products, while also providing technical thought leadership around APIs and microservices within both Kong and the external software community. Prior to Kong, Marco co-founded Mashape in 2010, which became the largest API marketplace and was acquired by RapidAPI in 2017.

// Related Links

Website: https://konghq.com/ https://www.youtube.com/watch?v=odpPVeQZjHU https://www.thestack.technology/the-big-interview-kong-cto-marco-palladino/

~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~

Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExplore

Join our Slack community [https://go.mlops.community/slack]

Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)]

Sign up for the next meetup: [https://go.mlops.community/register]

MLOps Swag/Merch: [https://shop.mlops.community/]

Connect with Demetrios on LinkedIn: /dpbrinkm

Connect with Marco on LinkedIn: /marcopalladino/

Timestamps:

[00:00] Agent-mediated interactions shift

[01:17] Kong connectivity and agents

[04:36] Transcript cleanup request

[08:11] MCP server use cases

[12:37] Agent world possibilities

[15:55] Business communication evolution

[18:55] System optimization

[25:36] AI gateway patterns

[31:30] Investment decision making

[35:54] Building conviction process

[41:34] Polished customer conversation

[46:37] AI gateway R&D future

[50:52] Wrap up

AI Conversations Powered by Prosus Group

Unicorn Mafia won the recent hackathon at Raise Summit and explained to me what they built, including all the tech they used under the hood to make their AI agents work.

Winners:

Charlie Cheesman - Co-founder at 60x.ai

Marissa Liu - Tech Lead, Reporting at Watershed

Ana Shevchenko - Software Engineer II at Spotify

Fergus McKenzie-Wilson - Co-founder at 60x.ai

Alex Choi - Founding Engineer at Medfin

Host:

Demetrios Brinkmann - Founder of MLOps Community

~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~

Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExplore

Join our Slack community [https://go.mlops.community/slack]

Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)]

Sign up for the next meetup: [https://go.mlops.community/register]

MLOps Swag/Merch: [https://shop.mlops.community/]

On-Device AI Agents in Production: Privacy, Performance, and Scale // MLOps Podcast #340 with NimbleEdge's Varun Khare, Founder/CEO and Neeraj Poddar, Co-founder & CTO.

Join the Community:

https://go.mlops.community/YTJoinIn

Get the newsletter: https://go.mlops.community/YTNewsletter

// Abstract

AI agents are transitioning from experimental stages to performing real work in production; however, they have largely been limited to backend task automation. A critical frontier in this evolution is the on-device AI agent, enabling sophisticated, AI-native experiences directly on mobile and embedded devices. While cloud-based AI faces challenges like constant connectivity demands, increased latency, privacy risks, and high operational costs, on-device breaks through these trade-offs.

We'll delve into the practical side of building and deploying AI agents with “DeliteAI”, an open-source on-device AI agentic framework. We'll explore how lightweight Python runtimes facilitate the seamless orchestration of end-to-end workflows directly on devices, allowing AI/ML teams to define data preprocessing, feature computation, model execution, and post-processing logic independently of frontend code. This architecture empowers agents to adapt to varying tasks and user contexts through an ecosystem of tools natively supported on Android/iOS platforms, handling all the permissions, model lifecycles, and many more.

// Bio

Varun Khare

Varun is the Founder and CEO of NimbleEdge, an AI startup pioneering privacy-first, on-device intelligence. With an academic foundation in AI and neuroscience from UC Berkeley, MPI Frankfurt, and IIT Kanpur, Varun brings deep expertise at the intersection of technology and science. Before founding NimbleEdge, Varun led open-source projects at OpenMined, focusing on privacy-aware AI, and published research in computer vision.

Neeraj Poddar

Neeraj Poddar is the Co-founder and CTO at NimbleEdge. Prior to NimbleEdge, he was the Co-founder of Aspen Mesh, VP of Engineering at Solo.io, and led the Istio open source community. He has worked on various aspects of AI, networking, security, and distributed systems over the span of his career. Neeraj focuses on the application of open source technologies across different industries in terms of scalability and security. When not working on AI, you can find him playing racquetball and gaining back the calories spent playing by trying out new restaurants.

// Related Links

Website: https://www.nimbleedge.com/

https://www.nimbleedge.com/blog/why-ai-is-not-working-for-you

https://www.nimbleedge.com/blog/state-of-on-device-ai

https://www.youtube.com/watch?v=Qqj_Nl2MihE

https://www.linkedin.com/events/7343237917982527488/comments/

~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~

Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExplore

Join our Slack community [https://go.mlops.community/slack]

Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)]

Sign up for the next meetup: [https://go.mlops.community/register]

MLOps Swag/Merch: [https://shop.mlops.community/]

Connect with Demetrios on LinkedIn: /dpbrinkm

Connect with Varun on LinkedIn: /vkkhare/

Connect with Neeraj on LinkedIn: /nrjpoddar/

Timestamps:

[00:00] On-device AI skepticism

[02:47] Word suggestion for AI

[06:40] Optimizing unique challenges

[13:39] LLM on-device challenges

[20:34] Agent overlord tension

[23:56] AI app constraints

[29:23] Siri limitations and trust gap

[32:01] Voice-driven app privacy

[35:49] Platform lock-in vs aggregation

[42:26] On-device AI optimizations

[45:38] Wrap up

AI Conversations Powered by Prosus Group 

Your AI agent isn’t failing because it’s dumb—it’s failing because you refuse to test it. Chiara Caratelli cuts through the hype to show why evaluations—not bigger models or fancier prompts—decide whether agents succeed in the real world. If you’re not stress-testing, simulating, and iterating on failures, you’re not building AI—you’re shipping experiments disguised as products.

Guest speaker: Chiara Caratelli - Data Scientist @ Prosus Group

Host: Demetrios Brinkmann - Founder of MLOps Community

~~~~~~~~ ✌️Connect With Us ✌️ ~~~~~~~

Catch all episodes, blogs, newsletters, and more: https://go.mlops.community/TYExplore

Join our Slack community [https://go.mlops.community/slack]

Follow us on X/Twitter [@mlopscommunity](https://x.com/mlopscommunity) or [LinkedIn](https://go.mlops.community/linkedin)]

Sign up for the next meetup: [https://go.mlops.community/register]

MLOps Swag/Merch: [https://shop.mlops.community/]