Jithendranath Joijoide

💥Mentorship and collaboration opportunity on MidcurveLLM is now open. 💥 ➡️ This project focuses on fine-tuning LLMs for geometric midcurve generation. ➡️ You will work on non-NLP, geometry-centric datasets using text-based formats. ➡️ The goal is to generate midcurves from 2D closed shapes with high precision. ➡️ Learn why fine-tuning works better than RAG for geometry problems. ➡️ Get hands-on with LLM fine-tuning, transformers, and benchmarking. ➡️ Ideal for practitioners with LLM fine-tuning and Python experience. ➡️ This is a voluntary, research-focused learning opportunity. Read the full post here: https://lnkd.in/dbXNgYfu #spatialai #AI #LLM #FineTuning #GeometryAI #Research #OpenSource

29

Baseten is tackling some of the most complex challenges in AI. Scaling inference is fundamentally a systems and infrastructure problem, not just an AI or model problem. As AI adoption grows, the real differentiator becomes the systems work that makes models reliable, efficient, and fast in production. By adopting NVIDIA's powerful open-source software and frameworks, we’re able to solve these challenges on behalf of our customers and deliver high-performance inference at scale. To learn more, check out our inference whitepaper 👇

20

1 Comment

Murmell joins the NVIDIA Inception program 🚀 Building collaborative, persistent cloud infrastructure for parallel AI coding agents requires serious compute, low-latency orchestration, and rock-solid scalability. Joining NVIDIA Inception gives us direct access to: World-class GPU infrastructure and technical support Advanced compute resources to scale our cloud sandboxes Direct connections within the global AI and startup ecosystem When we started building Murmell, our goal was clear: take multi-agent workflows out of local terminal silos and move them into persistent, multi-provider cloud environments where teams and agents code together without merge conflicts. Having NVIDIA back our journey validates the architectural direction we're taking. Huge milestone with Moss'Ab Mirande-Ney. We're just getting started. ⚡️ #NVIDIA #NVIDIAInception #AI #DevTools #CloudComputing #MultiAgentSystems #BuildInPublic #Startups

32

7 Comments

inference is eating ai. 40%+ of nvidia’s revenue now comes from inference workloads.. jensen huang says it’ll scale “a billion times” as chain-of-thought models go mainstream. the bottleneck isn’t chips anymore. it’s watts….performance-per-watt is the governor. huang claims even if rivals gave GPUs away for free, customers would still pick nvidia because blackwell = 30x more tokens per watt than hopper. what does this mean for the rest of us? 1) if you’re building infra-adjacent: efficiency is your narrative. every buyer cares about energy-per-outcome. 2) sellers need to frame solutions as part of ai factories.. not point tools. buyers want orchestration + integration, not another silo. 3) investors: watch inference-heavy categories and sovereign ai buildouts. this is where the multi-trillion upside lives. plus>> when the c-suite realizes watts = revenue, every board deck will start looking more like an energy strategy than a product roadmap. this is take place as more workflows start to move from human capital to agents — ps > more of this on my substack.. link in comments.

12

3 Comments

NVIDIA’s $12.93B agreement to acquire Hugging Face puts model choice in the spotlight. My take: access to models is only the beginning. The real differentiator is what they deliver when the demo ends and the customer starts talking. That’s why I’m proud to share a double milestone for Deepdub: 🚀 Phantom Z 3.4 Conversational is live. Our latest multilingual text-to-speech model brings expressive voice, real-time performance, full 48 kHz audio, and improved handling of the numbers, dates, and amounts that matter in customer conversations. 🏆 Our emotional TTS model ranked #1 for p95 latency among the expressive models in the Coval platform comparison. Expressive speech backed by a measurable speed result. Coval is the voice AI evaluation platform built to test and improve agents at scale. For me, these milestones belong together. Speed without expression isn’t enough. Expression that arrives too late doesn’t feel like a conversation. A worried customer should hear reassurance. An account number should be spoken clearly. And nobody should have to choose between a responsive agent and a natural-sounding one. NVIDIA’s announcement reinforces my conviction: the opportunity isn’t just to make more models available. It’s to build specialized models that earn their place in production. That’s what we’re building at Deepdub. Proud of the team for pushing both performance and expression forward. The next generation of AI agents won’t just need the right words. They’ll need the right voice. https://lnkd.in/dRYapGMP #Deepdub #VoiceAI #ConversationalAI #NVIDIA

10

Most teams running LLM inference are over-provisioning GPUs “just in case.” We benchmarked what happens when you combine NVIDIA NIM with NVIDIA Run:ai with the same inference workloads on dramatically fewer GPUs — with minimal throughput loss. How? ✅ GPU Fractions + Bin Packing: Run multiple NIM models on shared GPUs with true memory isolation — consolidate your deployments, increase density, and reduce your compute footprint. ✅ Dynamic GPU Fractions: Let models burst into unused headroom during traffic spikes — no extra GPUs required. ✅ GPU Memory Swap: Keep infrequently used models warm in CPU memory to slash first-request latency. ✅ Inference Priority Scheduling: Prioritize user-facing inference while training flexes around it. The takeaway: You may not need more GPUs — just smarter orchestration of the ones you have. 🔗 Full benchmarks and methodology in the tech blog: https://bit.ly/407QnUs

3

NVIDIA’s planned acquisition of Hugging Face is interesting to me because it shows how the AI battle is moving beyond just models. NVIDIA is already strong in AI compute, while Hugging Face has built a strong ecosystem around models, datasets and AI development. The AI stack is becoming more important end to end: Compute → Models → Model Registry → Harness/Runtime → Agents → Enterprise Applications. For enterprises, I think the key takeaway is to avoid dependency at every layer, not just the LLM layer. We should build AI architecture with portability and interoperability in mind, so we can change models, tools and platforms as technology continues to evolve. #EnterpriseAI #AIArchitecture #AgenticAI #NVIDIA #HuggingFace

5

Quick week in review on AI compute - Sept 2026: – Lead times still brutal for most buyers – Pricing spreads between neo clouds and hyperscalers remain huge – Power and transformers continue to constrain new builds – Private / dedicated capacity is solving the problem for a growing number of teams We run private GPU infrastructure single or multi tenant and help source hardware. If you’re dealing with any of the above, reply “private” or DM me. I’ll respond personally.

2

Some of Nvidia's best engineers have stayed with Jensen Huang for 30+ years. That's an unusual kind of retention in tech. Nvidia went through near-failure in its early years. The product struggled, the market was uncertain, and the company had every reason to become more controlling. Huang chose a different approach: give talented people ownership instead of managing them through fear. That philosophy scaled with the company. → Engineers were trusted to solve problems independently. → Senior talent stayed long enough to build deep institutional knowledge. → The company grew without replacing its engineering culture with layers of control. 3 leadership lessons: 1️⃣ Give your best people room to operate. High performers rarely need someone watching every decision. They need the authority to make them. 2️⃣ Retention is often a byproduct of trust. People don't stay for decades because of one great incentive. They stay when they consistently feel trusted, challenged and valuable. 3️⃣ Scale shouldn't mean more control. A common leadership mistake is becoming more bureaucratic as the company grows. Huang's case shows the opposite: if autonomy helped attract great talent early, taking it away later can be exactly what makes that talent leave. The real competitive advantage isn't just hiring exceptional people. It's building a company where exceptional people still want to be there decades later.

343

31 Comments