Episodi

  • $20M to Fix AI's Most Expensive Problem | Ep 09 (Junchen Jiang CEO, Tensormesh)

    $20M to Fix AI's Most Expensive Problem | Ep 09 (Junchen Jiang CEO, Tensormesh)
    Jul 14 2026

    👨🏽‍🚀 Welcome to Episode 10 of "Tech Beats Unplugged" with Junchen Jiang, CEO & Co-Founder of Tensormesh.


    $20M to Fix AI's Most Expensive Problem.

    Fresh off Tensormesh's $20M funding announcement backed by AMD Ventures, CoreWeave, NVentures (NVIDIA), Valley Capital Partners, and Laude Ventures, Junchen joins Kosseila HD to unpack the journey from a research insight at the University of Chicago to building one of the fastest-growing AI infrastructure companies.


    In this episode, we dive into the story behind LMCache, why KV cache evolved from a largely misunderstood concept in academia into one of the most critical layers of modern AI inference, and how Tensormesh is building an open, caching-accelerated inference platform for the next generation of AI.


    We're excited to share one of our most technical conversations yet.

    Topics discussed:

    1. (00:00) Introduction
    2. (01:08) The Origin Story and Insight Behind LMCache
    3. (06:27) The Origin of "KV Cache" in Academia
    4. (11:25) Day 2 of AI: The Inference Bottleneck
    5. (14:59) The Secret to Sustaining a Thriving Open Source Community
    6. (16:42) Tensormesh Inference Platform V1 and $0 Cached Tokens
    7. (18:28) Designing an Open, Agnostic Ecosystem (Serverless to On-Prem)
    8. (21:42) Working with Hardware Vendors
    9. (24:01) KV Cache in the Agentic Era
    10. (27:38) What GPU & AI Cloud Backing Says About Tensormesh
    11. (29:49) Closing Thoughts: Where Tensormesh Goes Next

    Show Notes

    • LMCache (Open Source): https://github.com/LMCache/LMCache
    • Tensormesh: https://www.tensormesh.ai/
    • Tensormesh Funding Announcement: https://www.tensormesh.ai/blog-posts/
    • Follow Tensormesh on X: https://x.com/tensormesh
    • Connect with Junchen Jiang: https://www.linkedin.com/in/junchenjiang/


    🎙 About Junchen Jiang

    Junchen Jiang is the CEO & Co-Founder of Tensormesh and the co-creator of LMCache, an open-source project developed from research at the University of Chicago. His work focuses on large-scale AI inference infrastructure, KV cache systems, and building an open ecosystem that spans GPUs, inference engines, storage, and enterprise AI deployments.


    Follow Tech Beats Unplugged for more conversations with founders, researchers, and engineers building the future of AI infrastructure, open source, cloud, and startups.
    Mostra di più Mostra meno
    35 min
  • 𝗧𝗵𝗲 𝗜𝗻𝗴𝗹𝗼𝗿𝗶𝗼𝘂𝘀 𝗖𝗹𝗼𝘂𝗱 𝗧𝗿𝗮𝗽𝘀 ☁️| Ep 09

    𝗧𝗵𝗲 𝗜𝗻𝗴𝗹𝗼𝗿𝗶𝗼𝘂𝘀 𝗖𝗹𝗼𝘂𝗱 𝗧𝗿𝗮𝗽𝘀 ☁️| Ep 09
    Feb 17 2026
    👨🏽‍🚀 Welcome to Episode 09 of "Tech Beats unplugged" 𝗜𝗻𝗴𝗹𝗼𝗿𝗶𝗼𝘂𝘀 𝗖𝗹𝗼𝘂𝗱 𝗧𝗿𝗮𝗽𝘀 ☁️" With Mark Boost CEO at Civo Cloud .𝗜𝘀 𝘁𝗵𝗲 𝗖𝗹𝗼𝘂𝗱 𝗯𝗿𝗼𝗸𝗲𝗻 ...𝗼𝗿 𝗷𝘂𝘀𝘁 𝘂𝗻𝗳𝗮𝗶𝗿? This time, Mark Boost and Kosseila H. go deep into the infamous 𝗖𝗹𝗼𝘂𝗱 𝘁𝗿𝗮𝗽𝘀 and explore where control is lost - and how to regained it for customers. 🌟 We’re so excited to share our latest tech Beats show with you! Topics discussed: (00:23) Introduction(01:42) From Formula 3 to serial entrepreneur (08:44) lessons from s𝗰𝗮𝗹𝗶𝗻𝗴 𝗺𝘂𝗹𝘁𝗶𝗽𝗹𝗲 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 (pivoting, competition, timing) (18:59) CIVO Elevator pitch (21:50) Why the 𝗰𝗹𝗼𝘂𝗱, 𝗮𝘀 𝘄𝗲 𝗸𝗻𝗼𝘄 𝗶𝘁, 𝗶𝘀 𝗯𝗿𝗼𝗸𝗲𝗻 ? (28:59) The infamous cloud traps: egress fees, exit tax (37:50) startup FREE credit : what's the catch? (44:02) Cloud regulation & fairness: blessing or curse? (47:32) Is Regulation killing AI innovation in Europe ? (51:31) David vs Goliath: independent providers vs the Big 3 hyperscalers "the MOAT"(56:56) 𝗔𝗜 𝗲𝗻𝗲𝗿𝗴𝘆 𝗰𝗿𝗶𝘀𝗶𝘀: how do we power AI without torching the grid👨🏻‍🔧🏭?(01:01) ESG vs Greenwashing? (01:03) 𝗢𝘂𝘁𝗮𝗴𝗲 𝗰𝗵𝗮𝗼𝘀: CEO mistakes — and how to do it right (CrowdStrike case)𝗦𝗵𝗼𝘄 𝗡𝗼𝘁𝗲𝘀Civo website : civo.comEgress Fees: Big Tech Braces For Digital Market Act Civo talk "The cloud is broken": The Cloud is Broken Digital Market Acts ⁠European Data Act⁠⁠Free credits:A Call for Fair Competition by Addressing the Impact of Free Cloud CreditsGreen policies ESG: ESG reporting: Carbon in the cloudData Center Co2 footprint calculator🎙𝗔𝗯𝗼𝘂𝘁 𝗠𝗮𝗿𝗸 𝗕𝗼𝗼𝘀𝘁:⁠⁠⁠⁠⁠⁠⁠⁠Website: ⁠⁠https://markboost.com/ LinkedIn: ⁠⁠ ⁠markboost X(Twitter): ⁠@markboost10 tips for a successful career
    Mostra di più Mostra meno
    1 ora e 11 min
  • 𝗡𝘂𝘁𝗮𝗻𝗶𝘅 𝗶𝗻 𝘁𝗵𝗲 𝗔𝗴𝗲 𝗼𝗳 𝘁𝗵𝗲 𝗩𝗶𝗿𝘁𝘂𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝗦𝗵𝗮𝗸𝗲-𝗨𝗽 (w/ Michael webster) | Ep08

    𝗡𝘂𝘁𝗮𝗻𝗶𝘅 𝗶𝗻 𝘁𝗵𝗲 𝗔𝗴𝗲 𝗼𝗳 𝘁𝗵𝗲 𝗩𝗶𝗿𝘁𝘂𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝗦𝗵𝗮𝗸𝗲-𝗨𝗽 (w/ Michael webster) | Ep08
    Dec 23 2025

    👨🏽‍🚀 Welcome to Episode 08 of "Tech Beats unplugged"

    🌐🦘This time the episode cuts through the 𝗩𝗠𝘄𝗮𝗿𝗲 𝗕𝗿𝗼𝗮𝗱𝗰𝗼𝗺 chaos and breaks down where Nutanix really fits in today’s virtualized/multicloud world. With the help of no other than Nutanix Guru Michael Webster all the way from News Zeeland

    🎧 We’re so excited to share our latest tech Beats show with you🧡! Please share away 🤗

    We hope you'll enjoy it!!!

    Topics discussed:

    1. (00:21) Introduction
    2. (01:58) What's your story with Nutanix
    3. (09:41) Hyper Converged Infrastructure. What's that?
    4. (13:38) Understanding Nutanix.
    5. (18:25) The VMware–Broadcom Chaos for customers
    6. (26:00) Running VMware on Nutanix (and moving off)
    7. (30:05) Nutanix hardware ... a Lego of Nodes
    8. (36:20) Nutanix building blocks - What's Controller VM
    9. (43:05) Nutanix 𝗸𝗶𝗹𝗹𝘀 𝘁𝗵𝗲 𝗹𝗲𝗴𝗮𝗰𝘆 𝗥𝗔𝗜𝗗 storage - how?
    10. (52:04) Nutanix; from 𝗵𝗮𝗿𝗱𝘄𝗮𝗿𝗲 𝘁𝗼 𝘀𝗼𝗳𝘁𝘄𝗮𝗿𝗲 company
    11. (57:00) What is Nutanix MOAT going forward ?


    Show Notes

    • VMware customers face up to 1,500% price increase News⁠

    • ⁠Nutanix How it works: ⁠⁠⁠Video

    • Dutch court forces Broadcom to support VMware migration after 85% price hike: ⁠Networkworld⁠

    • Nutanix Bible (nutanixbible.com)⁠

    • ⁠AT&T sues broadcom over licensing: ⁠⁠⁠⁠Forbes

    • Nutanix AHV: ⁠Built-in Enterprise-Class Virtualization⁠

    🎙About Michael Webster:

    • ⁠⁠⁠⁠⁠⁠Website: ⁠longwhiteclouds.com⁠

    • LinkedIn: ⁠Steve Giguere⁠

    • X(Twitter): @vcdxnz001


      Brought to You by Cloudthrill.

    Mostra di più Mostra meno
    1 ora e 3 min
  • 🔴TechBeats live : LLM Quantization "vLLM vs. Llama.cpp" | Ep07

    🔴TechBeats live : LLM Quantization "vLLM vs. Llama.cpp" | Ep07
    Jul 19 2025

    👋🏼 Hey AI heads 🎙️
    Join us for the very first Tech Beats Live 🔴, hosted by Kosseila—aka @CloudDude from @CloudThrill.

    🎯 This chill & laid-back livestream will unpack LLM quantization 🔥:

    • ✅ WHY it matters
    • ✅ HOW it works
    • ✅ Enterprise (vLLM) vs Consumer (@Ollama) trade-offs
    • ✅ and WHERE it’s going next.

    We’ll be joined by two incredible guest stars to talk Enterprise vs Consumer Quantz 🗣️:

    🔷 Eldar Kurtić – bringing the enterprise perspective with vLLM.
    🔷 Colin Kealty – aka Bartowski, creator of the top-downloaded GGUF quantized LLMs on Hugging Face.

    🫵🏼 Come learn and have some fun 😎.

    𝐂𝐡𝐚𝐩𝐭𝐞𝐫𝐬:

    (00:00) Host Introduction
    (04:07) Eldar Intro
    (07:33) Bartowski Intro
    (13:04) What’s Quantization!
    (16:19) Why LLM Quantization Matters?
    (20:39) Training vs Inference – “The New Deal”
    (27:46) Biggest Misconception About Quantization
    (33:22) Enterprise Quantization in Production (vLLM)
    (48:48) Consumer LLMs & Quantization (Ollama, llama.cpp, GGUF) – “LLMs for the People”
    (01:06:45) BitNet 1-Bit Quantization from Microsoft
    (01:28:14) How Long It Takes to Quantize a Model (Llama-3 70B) – GGUF or lm-compressor
    (01:34:23) What Is I-Matrix & Why People Confuse It with IQ Quantization?
    (01:39:36) What’s LoRA & LoRA-Q?
    (01:42:36) What Is Sparsity?
    (01:47:42) What Is Distillation?
    (01:52:34) Extreme Quantization (Unsloth) of Big Models (DeepSeek) at 2-bits 70 % Size Cut
    (01:57:27) Will Future Models (Llama-5) Be Trained on FP4 Tensor Cores?
    (02:02:15) The Future of LLMs on Edge Devices (Google AI Edge)
    (02:08:00) How to Evaluate the Quality of a Quantized Model
    (02:26:09) Hugging Face’s Role in the World of LLM/Quantization
    (02:33:46) Hugging Face’s Role in the World of LLM/Quantization
    (02:36:41) LocalLlama Sub-Reddit Down (Moderator Goes Bananas)
    (02:40:11) Guests’ Hope for the Future of LLMs & AI in General

    📖 Check out the quantization blog: https://bitly/LLMQuant

    #AI #LLM #Quantization #TechBeatsLive #LocalLlama #vLLM #Ollama
    Mostra di più Mostra meno
    2 ore e 51 min
  • Ep06: "GitHub Security horror stories " (with Steve Giguere)

    Ep06: "GitHub Security horror stories " (with Steve Giguere)
    Jun 10 2025
    👨🏽‍🚀 Welcome to Episode 06 of "Tech Beats unplugged" This time, we’re diving headfirst into 𝐭𝐡𝐞 𝐜𝐫𝐚𝐳𝐢𝐞𝐬𝐭 𝐆𝐢𝐭𝐇𝐮𝐛 𝐬𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐬𝐭𝐨𝐫𝐢𝐞𝐬, and who better to join us than Steve Giguere, an industry veteran and security expert who’s seen it all.From supply chain security mayhem to GitHub Actions gone wrong, we uncover real-world security blunders, attack vectors, and best practices to keep your repos and workflows safe.🌟 We’re so excited to share our latest tech Beats show with you🧡! Please share away 🤗We hope you'll enjoy it!!!Topics discussed: (00:00) Introduction(03:53) Software Supply Chain Security acronyms (SAST, DAST, IAST, etc.)(09:15) “A workflow is an application within your application” - What does that mean?!(12:16) Public vs. Private Repos - Are private orgs still at risk?(18:27) Self-hosted runners: Safe or security nightmare?(21:16) GitHub Environment Variables - How critical are they?(22:55) Secrets, masks, and how secure they really are (28:05) Artifact vs. Caching: Which is safer?(31:27) Craziest GitHub security screw-ups Steve has ever seen 🔥(36:42) Common attack vectors in GitHub Actions(44:19) Best security practices for GitHub Actions - Low-hanging fruit fixes 🍏(50:22) Are public actions safe? Can they be scanned?(53:52) xz backdoor fiasco - Lessons from the latest supply chain attack(59:00) NVD’s slowdown - What’s at stake?Show NotesCI/CD Goat (Deliberately vulnerable CI/CD environment): GitHubGitHub cache poisoning: Cacheract Attack | ScribeSecurityYour GitHub Secrets in Plain Text: CloudThrillGhat tool (Updating dependencies in GitHub Actions): GitHubOpenSSF Scorecard: WebsiteThe GitHub Worm (Asi Greenholts): Palo Alto BlogOWASP Top 10 CI/CD Risks: OWASPHeartbleed OpenSSL Exploit: Wikipedia🎙About Steve Giguere:⁠⁠⁠⁠Website: stevegiguere.comLinkedIn: Steve GiguereBook: Cloud Native Application Protection Platforms – O'ReillyPersonal Blog: CodifyreTalk Lessons Learned from OSS and GitOps Journey: YouTubeOWASP Lisbon Talk: YouTubeStayWiredIn YouTube Show: StayWiredInDevSecOps Podcast: Spotify
    Mostra di più Mostra meno
    1 ora e 6 min
  • Ep05: "Deploy Local LLMs 𝐢𝐧 the Cloud (𝟏𝟎𝟎% 𝐃𝐚𝐭𝐚 𝐏𝐫𝐢𝐯𝐚𝐜𝐲)"

    Ep05: "Deploy Local LLMs 𝐢𝐧 the Cloud (𝟏𝟎𝟎% 𝐃𝐚𝐭𝐚 𝐏𝐫𝐢𝐯𝐚𝐜𝐲)"
    Sep 24 2024

    👨🏽‍🚀 Welcome to Episode 05 of "Tech Beats unplugged"

    This time, we tried something completely crazy – we're letting the AI hosts take over! That's right 😎. We're flipping the script and giving the AI the mic to guide us through the fascinating world of local LLMs. but that's not all as this episode is actually inspired by my recent talk at Oracle Cloud World in Vegas. The topic? You guessed it: Local LLMs in the cloud.

    🌟 We’re so excited to share our latest tech Beats show with you🧡!

    We hope you'll enjoy it!!!

    Topics discussed:

    1. (00:00) Introduction
    2. (01:00) Why OpenAI Might Not Be Your BFF?
    3. (02:40) Local/Open LLMs to the Rescue!
    4. (03:38) What's Quantization!
    5. (04:30) Where to find these Open LLMs?
    6. (05:02) Inference Engines (Ollama)!
    7. (05:50) What's a modelfile!
    8. (06:40) What about deploying local AI to the cloud?(OKE/managed kubernetes)
    9. (07:30) From zero to cloud deployment Hero
    10. (08:28) What's Next (LLM ethic benchmark)
    11. (09:55) Outro.

    Show Notes

    • My local LLM GitRepo: Ollama_lab
    • Helm leaderboard for model safety: Sandford Helm model leaderboard
    • My talks in Oracle cloud world 2024: OCW2024LLM
    Mostra di più Mostra meno
    11 min
  • Ep04: "Networking in the Cloud Part 1" (with Jose Moreno)

    Ep04: "Networking in the Cloud Part 1" (with Jose Moreno)
    Apr 17 2024

    👨🏽‍🚀 Welcome to Episode 04 of "Tech Beats unplugged"

    This time, we dared to tackle the big, yet exciting topic which is "𝐍𝐞𝐭𝐰𝐨𝐫𝐤𝐢𝐧𝐠 𝐢𝐧 𝐭𝐡𝐞 𝐜𝐥𝐨𝐮𝐝", and to help us with this challenge we summoned our expert and networking GURU "Jose Moreno" all the way from Munich Germany.

    As an Azure Principal Engineer at Microsoft, he comes to share his knowledge and demystify all those mysterious networking concepts, answering questions you never dared to ask in the universe of SDN and cloud networking😀( SDWAN, Hub&Spoke, migrations, Cloud lies, you name it).

    🌟 We’re so excited to share our latest tech Beats show with you🧡!

    We hope you'll enjoy it!!!

    Topics discussed:

    1. (00:00) Introduction
    2. (04:00) How did you enter IT and the networking world ?
    3. (06:14) What has really changed for network admins since Cisco golden years ?
    4. (09:30) Core differentiators between on-prem and Cloud/SDN paradigm?
    5. (15:29) IPV6 Doom day: Why are we so stuck with IPV6 adoption?
    6. (21:58) What lab tools can students use to prep for cloud networking exams?
    7. (28:30) What the heck is BGP & OSPF and are they still used in the cloud?
    8. (38:09) Your cloud subnet constructs is fake😱😱???
    9. (46:50) Difference between WAN & SDWAN, how it integrates with the cloud ?
    10. (50:52) Hub & Spoke Versus Virtual WAN pros and cons?
    11. (56:54) Biggest networking challenge stories when moving to the cloud ?
    12. (1:00:43) I'm a small shop CTO moving Hybrid, help me out! #BestPractices ?
    13. (1:05:41) What's the hype around IPSEC tunnels over Direct connect?
    14. (1:12:9) Your most helpful contributions to the cloud networking community ?


    Show Notes

    • IPV6 & IPV4: ⁠⁠⁠⁠⁠⁠⁠two-internets-both-flakey by Avery Pennarun⁠
    • your cloud subnet constructs is fake: ⁠new-cloud-security-thinking(podcast)⁠
    • Roggers BGP outage: ⁠Summary⁠
    • ⁠BGP: ⁠BGP School for cloud people(podcast)⁠
    • VPN Aviatrix insane mode: Article

    Jose Moreno:

    • ⁠⁠⁠linkedin⁠
    • ⁠Blog⁠
    • ⁠Azure review checklists⁠⁠
    • ⁠⁠Github: erjosito/azure-networking-lab ⁠⁠
    • ⁠Microsoft Azure What The Heck k8 project⁠
    Mostra di più Mostra meno
    1 ora e 17 min
  • Ep03: "Journey into Tech with Kelsey Hightower (Uncut version) 🌠

    Ep03: "Journey into Tech with Kelsey Hightower (Uncut version) 🌠
    Nov 28 2023
    👨🏽‍🚀 Welcome to Episode 03 of "Tech Beats unplugged" This time, we have been incredibly honored and fortunate to host one of the most prolific speakers, storyteller, and thought leader of his generation. He was kind enough to bless us with his presence for an exclusive interview filled with priceless wisdom and life hacks💡. Beyond his groundbreaking contributions to Kubernetes, Cloud Native and Open-Source software, he's been one the most influential figures in the tech industry over the past decade. Today, despite his retirement, he continues to inspire engineers across the world to build smarter, scalable, and highly usable systems🌟. We’re so excited to share our privileged exchange with the one and only, Kelsey HighTower🧡! We hope you'll enjoy it!!! Topics discussed: (00:00) Introduction(02:06) Day 0: Kelsey's early life up to his tech introduction Fun facts, boyhood in Long BeachEarly hustles in Atlanta First gig in Tech (08:20) (14:40) If we had a dinner with 3 of your best friends and Asked them to describe you, What would we learn 🦸‍♂️ ? The past lives of Kelsey in the show business The impact of minimalism (25:40) What did it take on a hindsight: Skills, talent, hard work, luck or all of them? The revolution of the Open source mindsetCan do attitude, increase the luck surfaceThe most powerful language: "Human language" (32:30) The CNCF landscape: How did we get there? Limitless options; Is it a good or bad thing (37:50) Monolith vs Microservices debate: is it still a thing today? Stack Overflow & Prime Video examples (48:40) Platform engineering vs DevOps! What's that all about ? (53:38) At what time serverless stops making sense? Beyond cost and process length. (57:56) One day Kubernetes will die, what does life after K8 look like? (01:05:42) What the heck is service mesh/istio and is eBPF is here to replace it? keeping up with side cars vs ambient mesh and cilium. (01:14:16) Dev consumerism, are we in a disposable Dev language era? I'm a junior dev, should I pick the weekly hype or stick to my guns? (too many frameworks, APIs, runtimes). (01:23:32) How to distinguish Right Abstraction from the Bloat Abstraction? (1:27:20) Something called Empathetic engineering? Your experience back at google improving usability. (1:34:31) Tips for when you're obsessed with exciting new ideas but fail miserably to convey its potential to the higher ups? (1:44:30) Startup CTOs conundrum- should we keep it simple (monolith/managed) or go all-in shiny framework? (1:52:00) The state of the Black Community in Silicon Valley (2:05:07) Career wisdom for Youngbloods (2:11:50) Surprise Quiz😀: What Kind of Kelsey are you (Hip Hop version 🎵) Show Notes Monolithic success stories: (⁠Stack Overflow⁠ - ⁠PrimeVideo⁠)⁠⁠Kubecon Keynote: Reflection Kelsey.H⁠⁠The future of K8s, is that it has to go away(GitHub ReadMe) ⁠ Crunchbase news: ⁠⁠Black Founders continue to not be funded in Silicon Valley⁠⁠Total VC funding in the US raised by black founders(1.2%) ⁠⁠Black Employees at top tech companies (5%)⁠
    Mostra di più Mostra meno
    2 ore e 36 min