Nikhil Bansal

About Nikhil Bansal

I design and deliver scalable, customer-focused platforms that transform complex…

Experience

1. Software Development Engineer

   Intel

   May 2023 — Present · Santa Clara, CA, US

   Led PoC to production rollout of a container-based AI platform, delivering on-demand access to 30+ GPUs and Habana Gaudi nodes for training and inference of 25+ GenAI models, including RAG, LLMs, Stable Diffusion, image-to-image, text-to-image, DeepSeek, & PyTorch.Took full ownership of platform delivery, authoring design documentation, architectural diagrams, and recovery plans, while leading stakeholder/product team discussions (ABR/SAFE reviews) and engaging directly with customers.Delivered core cloud-native features (Load Balancer, NFS Dynamic Volume Provisioning, Azure AuthN/AuthZ, microservice for Sign-up, callback routing, rate limiting, coupon-based access), enabling secure multi-tenant workloads supporting 520+ concurrent users.Optimized runtime flexibility by designing Dockerfiles in AWS ECR and integrating dynamic pod image selection, improving flexibility.Enhanced platform reliability by integrating Prometheus/Grafana monitoring, Nessus security scanning, and enforcing Istio + Calico network policies, reducing security risks by 46.5% and ensuring enterprise-grade observability, high Availability, and fault Tolerance.Enabled users to land in custom Kubernetes containers with a GPU Card, persistent storage, and pre-installed model-specific dependencies with intelligent node and GPU card scheduling, reducing setup time by 36.88%.Orchestrated infrastructure for automated provisioning of compute instances using Slurm partitions and a microservice-driven cron job scheduler, enabling just-in-time allocation of Intel GPUs, Gaudi accelerators, and CPU resources with Azure AD–based secure access.Executed backend microservices with REST APIs and PostgreSQL database, streamlining resource provisioning and workload scheduling.Created Ansible playbooks to standardize deployments, reducing misconfigurations by 30%Built multi-user functionality by extending AuthZ/AuthN flows with Enterprise ID, modernizing DB schema and APIs for early adopters.

Education

• University of North Carolina at Charlotte

  Master's degree, Computer Science

• Rajiv Gandhi Prodyogiki Vishwavidyalaya

  Bachelor of Technology - BTech, Computer Science Engineering

  2017 — 2021