Siva Govindaraj Gounder

About Siva Govindaraj Gounder

Results-driven technology leader with deep expertise in AI infrastructure, distributed…

Experience

1. Tech Lead Ai Infra Platform

   Stealth AI Startup

   Apr 2023 — Present · Bellevue, WA, US

   0->1 building high scale GPU AI infra orchestration platform ( for training and inference) using Immersion Cooling, Slurm, RDMA, NCCL, UCX, libfabric, Infiniband(IBV_Verbs), RoCE V2, GPU Direct, OpenMPI using C++ CUDA and BF3/BMC/DPU• Built a large-scale, robust HPC infrastructure platform to provision CUDA, MOFED, NCCL, Slurm, Sunk, GPU and Network operators.• Collaborated closely with customers to debug training and inference workload issues, performance bottlenecks in GPUs, NCCL, and RDMA.• Led the observability stack, developing RDMA agents, GPU agents, and Host agents with policy enforcement and drift management and Active GPU Health Monitoring.• Developed a fault-prevention platform for XID errors and GPU failures, Single and Double-bit errors incorporating watchdog functionality to monitor training performance bottlenecks.• Highly skilled in debugging failures and performance issues in software and hardware stack, with deep knowledge of CUDA internals, NV kernel, NVSHMEM, distributed training frameworks, GPUDirect RDMA, IBVerbs, NCCL, UCX, MPI.• Experts in debugging issues from GPU firmware, BMC, NVIDIA kernel-mode, and user-mode drivers leveraging NVML and Out-of-Band (OOB) Metrics.• Implemented GPU Fault-Tolerance and Resilience Platform to intercept and tolerate faults with multiple remediation strategies for various failure types.• Developed datacenter management tools to monitor GPU failures, streamline the RMA process, and debug and identify root causes of performance degradations.

Education

• Coimbatore Institute of Technology

  Bachelor of Engineering, Computer Science and Engineering

Skills

• Blackberry
• Windows Mobile
• Apparel & Fashion
• Design Patterns
• Visual C++
• Uml
• Ios Development
• Mobile Device Management
• Win32
• Databases
• Smart Homes
• Mobile Applications
• Technology Evangelism
• Industry 4.0
• Objective-C
• Embedded Systems
• C++
• Swift
• Machine Learning
• Mobile Applications Management
• Android
• Soap
• Ios
• Object Oriented Design
• 6lowpan
• Java
• Connected Devices
• J2me
• Json
• Cloud Computing
• Connected Health
• Xml
• Zigbee
• Mobile Payments
• Enterprise Mobility
• .net Framework
• C#
• .net
• Multithreading
• Digital Transformation