Aug 2024 – Present · 1 yr 10 mos · Mumbai · On-site

• Technical Co-Founder and solo engineer behind a core-AI B2B SaaS platform. I own the entire engineering surface - engineering management, full-stack development, application architecture, UI/UX design, cloud infrastructure, and DevOps.
• Building a multi-tenant system with RBACs and enterprise functionality. Core R&D work centered on large language model evaluation: creating internal evaluations for benchmarking frontier LLMs against custom enterprise use cases, analyzing cost-latency-intelligence trade-offs, and building AI-enabled workflows that power the platform's core product.
• Provision and manage infrastructure across Cloudflare, Railway, Convex, AWS and GCP. Handle corporate IT administration (Microsoft 365/Google Workspace, domain, employee provisioning). Involved in B2B sales and customer support. End-to-end company setup from incorporation, IT provisioning, product development and infrastructure, recruiting, management and other operational groundwork.

Jun 2022 – Jul 2024 · 2 yrs 1 mo

• Worked in datacenter division, on Blackwell (RTX Pro 6000) emulation through production, Hopper, Ampere, Pascal generations of NVIDIA GPUs for feature implementation, bugs, performance enhancements, etc. Handled escalations and feature implementations for high value clients - Disney, Pixar, Google (GCP), Alibaba (AliCloud), RedHat. Implemented live migration support of vGPU enabled VMs on KVM hypervisor as part of GPU virtualization team. Live migration benchmarking automations and test suite for functional and performance regressions. Worked on and specialised in NVIDIA open source - nvidia-vgpu-vfio module, RHEL-KVM kernel, and QEMU/Libvirt compatibility. vGPU - GSP CUDA performance investigations. vGPU vfio-pci-core model verification. Extensive experience with release planning, QA, and enterprise support. Ownership of several projects and internal productivity tools. Conducted evaluation and benchmarking of AI workflows and tools in late 2023 / early 2024, presented benefits to management and advocated for internal adoption.

Jul 2021 – May 2022 · 10 mos

• Worked in GeForce NOW team in datacenter division. Maintenance of a linux based hypervisor fork for GeForce NOW infrastructure, patching kernel, QEMU, libvirt and security upgrades. Automated hypervisor performance benchmarking and comparative statistical analysis across hardware generations, reducing turnaround time by 95%+ (Weeks to hours). Developed iPXE based ISO installation software enabling automated, instant, error-free internal OS installations globally, reducing server provisioning time by 99%+ (hours to minutes). Created elaborate onboarding documentation for new engineers, reducing time to first contribution from ~3 months to under a month. Mentored 3 new team members, conducted introductory onboarding seminars for ~40 interns.

Jan 2021 – Jun 2021 · 5 mos

• Implemented ~20% performance improvements for virtualised GPU workloads in multi NUMA node config datacenter servers: auto NUMA assignment for vGPU host processes and cache coloring techniques on noisy neighbor scenarios in a multi-tenant system

May 2020 – Jul 2020 · 2 mos · Mumbai

• Built deployment standardisation pipelines using Jenkins for multiple internal applications for delivery across bare-metal Docker and Kubernetes (OpenShift) environments. Led sessions to onboard multiple business units and migrate to container platforms and CI/CD adoption.