# Overview Shreyas Mocherl attended KubeCon + CloudNativeCon India 2026 in Mumbai for the first time and experienced it both as an attendee and as a speaker. He co-presented a technical session on turning an NVIDIA DGX Spark into a self-hosted AI cluster and left the conference with new certifications, connections, and professional confidence.
# The talk: DGX Spark, Kubernetes, GPUs, and DRA The talk, "Run Your Own AI Cluster on a DGX Spark: Kubernetes, GPUs, and DRA," was co-presented with Janakiram MSV, Shreyas's father. The session covered building a Kubernetes cluster on DGX Spark hardware, exposing GPUs to workloads, using Dynamic Resource Allocation to schedule GPU workloads, and serving models on infrastructure you own and control. A recording of the talk is available on YouTube.
Rehearsals and AV checks mattered more than Shreyas expected. Facing a hall of that size was nerve-testing, but the presentation shifted into a technical conversation once it started.
# Pins, certifications, and the Golden Kubestronaut
He attributes passing the certifications to community-built, free resources: open documentation, community-run study groups, and platforms like KodeKloud. The pins serve as a tangible reminder of that path.
# People and the hallway track Shreyas emphasizes that the most valuable part of KubeCon is informal interaction. He highlights several people he met:
- Saiyam Pathak — became a friend after sustained conversations during the event.
# What surprised a first-timer
- The event is large but navigable, and the community is approachable.
- A speaker badge changes interactions: people approach with questions, ideas, and sometimes job leads.
- India's cloud native community showed high energy, deep technical questions, and many students and early-career engineers, suggesting significant regional momentum.
# What he took home Beyond badges and photos, Shreyas left with three concrete outcomes:
- Confidence to stand on a KubeCon stage again and to write another CFP.
- Relationships formed at the event that he expects to compound over time.
- Momentum toward projects involving GPUs, DRA, and AI on Kubernetes, reinforced by weeklong technical conversations.
# Practical next steps for readers