Pavan Kumar Dhulipalla
Leander
Summary
Senior infrastructure architect with 10+ years building and operating carrier‑grade, low‑latency content delivery platforms across distributed edge sites for Tier‑1 operators. Specialized in GPU‑accelerated AI infrastructure, high‑performance networking, InfiniBand, RDMA, and Kubernetes-based distributed systems. Architected multi-site deployments of UPF-based edge content delivery, achieving sub‑millisecond routing and high-throughput traffic handling. Hands-on experience with NVIDIA GPU clusters, NCCL behavior, InfiniBand fabrics, and distributed observability stacks. Recognized for leading complex PoCs, shaping architecture with customers and engineering teams, and enabling adoption of next-generation distributed compute platforms.
Overview
12
12
years of professional experience
1
1
Certification
Work History
Principal Technology Architect
Microsoft
Austin
11.2020 - Current
- Architected and operated a distributed UPF‑based content delivery fabric across multiple edge regions on Azure Operator Nexus, achieving sub‑millisecond routing and high-throughput traffic processing.
- Led PoCs for SMF‑UPF distributed architectures, including selective breakout, session anchoring, latency optimization, multi‑site failover design, and comparative performance benchmarking.
- Validated NVIDIA/Mellanox ConnectX NIC resiliency under high-throughput workloads, testing firmware behavior, PCIe resets, congestion scenarios, and fabric anomalies—feeding results into product standards.
- Delivered sub‑1.5s convergence during compute, fabric, and NIC failure events with zero session displacement at ~87 Gbps loads, enabling carrier-grade reliability for distributed edge workloads.
- Built observability stacks using Prometheus, Grafana, and ELK for real-time metrics on throughput, microbursts, session behavior, packet drops, and NIC/SR‑IOV counters.
- Developed Bicep IaC templates and Azure DevOps pipelines to automate SMF/UPF deployments across dozens of edge POPs, ensuring deterministic rollouts and repeatable lifecycle operations.
- Served as technical anchor between engineering, product, and customer teams—shaping architecture decisions, feature prioritization, rollout plans, and operational guidelines.
Resident Engineer
Affirmed Networks
Redmond
11.2015 - 10.2020
- Supported EPC core deployments across 3000+ hosts and thousands of VMs, optimizing performance, reliability, and lifecycle operations for virtualized network functions.
- Executed large-scale performance modeling using IXIA and Spirent, characterizing throughput, latency, and scalability behavior for control and user plane elements.
- Implemented elasticity and lifecycle management using VNFM platforms, enabling automated scale-in/out workflows for critical workloads.
- Troubleshot complex protocol flows (SIP, Diameter, GTP, RTP) using NETSCOUT, IXIA, Wireshark, ensuring stability of mobility, session management, and policy control functions.
Junior Network Engineer
ISL Tech Solutions
02.2014 - 10.2015
- Worked on VoWiFi and VoLTE solution validation at AT&T Labs, including architecture integration, interop testing, and multi-vendor mobility flows.
- Gained deep exposure to mobile core protocols (SIP, Diameter, GTP), QoS enforcement, policy control, and latency-sensitive service delivery.
- Used tools such as NETSCOUT, IXIA, and Spirent for performance, troubleshooting, and traffic simulation across IMS and EPC components.
Education
M.S. - Electrical Engineering
Missouri University of Science & Technology
B.Tech - Electronics & Communications Engineering
JNTU Kakinada
Skills
- NCCL benchmarking
- NVLink and NVSwitch expertise
- IB fabric tooling
- RDMA fundamentals
- GPUDirect RDMA knowledge
- 5G UPF-based CDN
- Low-latency routing strategies
- SMF-UPF proof of concepts
- Distributed load balancing
- NVIDIA/Mellanox ConnectX proficiency
- Firmware lifecycle management
- DevOps automation pipelines
Certification
- Certified Kubernetes Administrator (CKA)
- Azure Administrator (AZ‑104)
- Azure Network Engineer (AZ‑700)
- Azure Fundamentals (AZ‑900)
- AI Fundamentals (AI‑900)
Ai Gpu Infrastructure Projects
- Ran NCCL all-reduce performance tests on multi-GPU nodes (A100/H100/H200) under varying payloads, comparing intra-node NVLink vs PCIe paths and analyzing bandwidth trends.
- Used InfiniBand tools (ibdiagnet, iblinkinfo, ibroute, perfquery) to study topology, link errors, congestion, and routing behavior in test GPU clusters.
- Explored GPUDirect RDMA data paths and mapped interactions between CPU, GPU, and NIC for distributed training and remote storage scenarios.
Timeline
Principal Technology Architect
Microsoft
11.2020 - Current
Resident Engineer
Affirmed Networks
11.2015 - 10.2020
Junior Network Engineer
ISL Tech Solutions
02.2014 - 10.2015
M.S. - Electrical Engineering
Missouri University of Science & Technology
B.Tech - Electronics & Communications Engineering
JNTU Kakinada