Building real AI agents on a desktop AI supercomputer, and showing everything, including what breaks.

I'm Ronak, a cloud-native engineer turned enterprise AI practitioner. On this channel I build agentic AI hands-on on my NVIDIA GB10: running open models locally, wiring up agent frameworks like CrewAI, LangGraph and NemoClaw, and deploying NVIDIA Blueprints end to end.

What you'll find here:
→ Step-by-step builds of agents and multi-agent workflows
→ Blueprint build-alongs, from reference architecture to a working stack
→ Local inference deep dives: vLLM, quantized models, performance tuning
→ Framework comparisons with honest verdicts
→ The debugging nobody shows you: container networking, dependency conflicts, driver quirks

No hype and no slides-only explainers. Every video is something I actually built and ran.