Zeeshan M.

Experience

1. Senior Staff Engineer

   Qualcomm

   Aug 2013 — Present · Irvine, CA, US

   AI/ML Reference Model Development- Designed and implemented a PyTorch-based AI/ML reference API optimized for Qualcomm NPU hardware, including custom bit-exact operations, Quantization-Aware Training (QAT) models, and full autograd support. Integrated new hardware instructions aligned with Neural Processing Runtime (NPR) standards and validated model performance on constrained edge devices.Cloud Infrastructure as Code Automation- Architected and automated Azure infrastructure using Terraform and Azure DevOps, implementing secure and scalable cloud solutions with IaC. Deployed PyTorch ML models via Azure ML pipelines, optimizing GPU-based workloads. Built CI/CD pipelines, developed secure Flask APIs, and migrated AI applications to Azure Databricks. Managed Azure services (AKS, SQL DB, Cosmos DB) within a hub-and-spoke architecture, while enforcing security compliance and leading enterprise cloud migrations.LTE/5G Modem DSP Software Development- Designed and developed 5G physical layer DSP firmware for channel estimation (DMRS, PTRS, CSIRS, TRS, Noise) and MU-MIMO detection on vector processors using Assembly/C. Contributed to vector processor evolution through instruction enhancements for performance gains. Built a test framework and verified microkernels across simulation and RTL platforms. Developed scheduler and interface firmware, supported chip bring-up and provided end-to-end integration on emulation platforms.