Jeff Lu

Learn to build conversational AI voice agents in "Building AI Voice Agents for Production", created in collaboration with LiveKit and RealAvatar, and taught by Russ d'Sa (Co-founder & CEO of LiveKit), Shayne Parmelee (Developer Advocate, LiveKit), and Nedelina Teneva, PhD (Head of AI at RealAvatar, an AI Fund portfolio company). Voice agents combine speech and reasoning capabilities to enable real-time conversations. They're already being used to support customer service, to improve accessibility in healthcare, for entertainment applications, and for talk therapy. In this course, you’ll learn to build voice agents that listen, reason, and respond naturally. You’ll follow the architecture used to create the "AI Andrew" Avatar, a collaborative project between DeepLearning.AI and RealAvatar that responds to users in what sounds like my voice. You’ll build a voice agent from scratch and deploy it to the cloud, enabling support for many simultaneous users. What you’ll learn: - Understand the fundamentals of voice agents, including key components like speech-to-text (STT), text-to-speech (TTS), and LLMs, and how latency is introduced at each layer. - Explore voice agent architectures and the trade-offs between modular pipelines and speech-to-speech APIs. - Explore how platforms like LiveKit mitigate latency issues with optimized networking infrastructure and low-latency communication protocols. - Learn how to connect client devices to voice agents using WebRTC—and why it outperforms HTTP and WebSocket for low-latency audio streaming. - Incorporate voice activity detection (VAD), end-of-turn detection, and context management to detect turns, handle interruptions, and manage conversational flow. - Understand the trade-offs between latency, quality, and cost in an example in which you build a voice agent and change its voice. - Equip your agent with metrics to measure latency at each stage of the voice pipeline and learn the key levers you can pull to make your agent faster and more responsive. The voice agents built in this course also incorporate voice technology from ElevenLabs, a supporting contributor to the project. By the end of this course, you'll have learned the components of an AI voice agent pipeline, combined them into a system with low-latency communication, and deployed them on cloud infrastructure so it scales to many users. I’m looking forward to seeing what voice agents you build from this course! Please sign up here: https://lnkd.in/gEBXAGH8

3,201

156 Comments

🚀 Vellum was featured in NVIDIA's recent Llama Nemotron Ultra launch blog post! We're impressed with the model's performance on GPQA Diamond, AIME benchmark and even our proprietary adaptive reasoning benchmark. It's fantastic to see the chip maker come out with SOTA open source models that can be used instead of closed source models for production use cases at a fraction of the cost. You can see how the model stacks up to other models on our LLM Leaderboard (currently top ranking leaderboard in the market) and learn more about it on their blog. LLM Leaderboard: https://lnkd.in/dqzT55ME NVIDIA launch announcement: https://lnkd.in/d9puXDUk

91

1 Comment

🔥 Jensen Huang might be the most unconventional CEO in tech — and that’s exactly why Nvidia is now a $3 trillion powerhouse. No 1:1s. No status reports. No 5-year plans. Yet he’s leading one of the most influential companies of our time. Here’s how he’s rewriting the leadership playbook 👇 🧠 Flat org, no private meetings 40 direct reports — no 1:1s Everything happens in groups “They already made it. They’re doing great” — no career coaching 📬 No status reports — just raw truth Doesn’t trust filtered updates Reads ~100 “top 5” emails daily from employees Wants unpolished, on-the-ground insight 🔄 Radical transparency No VP-only meetings Strategic decisions shared openly Everyone has context, all the time 🚫 No fixed plans No 1-year or 5-year plans Always adapting — especially in the fast-evolving world of AI This model is built to: ✔️ Attract exceptional talent ✔️ Spread information quickly ✔️ Scale fast without the bloat 👀 In a world full of rigid org charts and over-optimization, maybe the future of AI belongs to the CEO bold enough to ignore the rules. Agree or disagree — would this work at your company? #Leadership #AI #JensenHuang #Nvidia #OrgDesign #FutureOfWork #StartupCulture

48

2 Comments

NVIDIA Omniverse Isaac Sim 5.0 now lets us easily simulate critical warehouse incidents like toppling boxes, fires, and liquid spills—all within a highly realistic, physics-driven environment. This means you can now model not just the routine operations, but also rare and hazardous events that could impact safety, efficiency, and robot reliability. Why does this matter? - End-to-End Warehouse Simulation: Simulate everything from normal robot pick-and-place to unexpected emergencies, ensuring your robotics solutions are robust and ready for the real world. - Risk-Free Testing: Test how your robots and workflows respond to incidents, without risking equipment or safety. - Better AI Training: Generate rich, synthetic data for AI models, including rare edge cases that are hard to capture in real life. - Faster Iteration: With AuraML hosted platform, you get all the power of Isaac Sim—no setup, no DevOps headaches, just seamless, scalable cloud simulation. At AuraML, we’re building a fully-hosted platform for warehouse and robotics operations simulation. Our goal is to empower you to test, validate, and optimise your warehouse automation end-to-end, including those “what if” scenarios, without the usual friction. Ready to see how your robots handle the unexpected? Let’s talk! #Robotics #Simulation #IsaacSim #WarehouseAutomation #AI #AuraML #NVIDIA

141

6 Comments

Computer-Aided Engineering ( #CAE ) was an important part of NVIDIA's keynote at CES this week. To quote Jensen Huang, NVIDIA's CEO : “Everybody wants to digitalize the future of industries… There is so much waste, so much opportunity for automation in that $50 TRILLION of the world’s GDP.” That message rings true for many of us in the tech space. But there was one thing was missing from this interesting keynote… Robots, autonomous vehicles, and automated factories - 3 examples given by Jensen in this talk - all have one thing in common (aside from NVIDIA, of course). They are all rendered useless without #RF connectivity. RF connectivity is created w/ a variety of hardware  #antenna,  #radar, #wifi and #bluetooth #radio transmitters, #GPS and so on. And in order for digitalization to occur as Jensen described here, simulation of electromagnetics and RF that supports the development of ALL that RF hardware is a must! While developing #simulation software that delivers full fidelity, very accurate simulation results quickly is no small feat, Nullspace, Inc. is doing just that.    To listen to the #CAE highlight of the keynote: https://lnkd.in/gUxeMDnK And if you're ready to implement modern #EM simulation tools in your workflow, schedule a demo with our team: https://lnkd.in/gYYq6gJM * Logos in the background of the keynote image are of Ansys, Altair and Cadence - #CAE leaders in the engineering simulation space.

393

8 Comments

NVIDIA 's CEO, Jensen Huang, imagines AI factories running 24/7, like power plants, producing AI models to meet growing demand. He also talks about "digital humans"—AI-driven avatars that could transform customer service and entertainment. While ambitious, these ideas raise concerns about energy use and environmental impact. Looks like we'll need more power!

143

33 Comments