Ovii Job Board

Senior Performance Engineer, Intel Stack

Sarvam

Bengaluru, India • Onsite - Bengaluru, India • Full-Time • 5+ years

Posted 2026-06-04 Tech & Engg

Apply on employer site

Job Description

Ovii's Interpretation of the Role

Senior Performance Engineer driving end‑to‑end Intel inference stack performance for Sarvam's AI platform. Owns OpenVINO build, model quantization, and CPU/GPU/NPU optimization to meet strict SLAs.

Role Snapshot

  • Senior performance engineer
  • Intel inference stack specialist
  • OpenVINO build & quantization lead
  • CPU/GPU/NPU optimization
  • Edge model deployment
  • Infrastructure team member

Must-Have Requirements

  • ML deployment
  • Intel inference stack (OpenVINO)
  • OpenVINO model conversion & quantization
  • ONNX Runtime EP knowledge
  • x86 CPU profiling & optimization
  • Intel inference stack experience

Nice-to-Have Signals

  • AVX‑512 or AMX intrinsics
  • Custom OpenVINO operator authoring
  • Direct prior interaction with OpenVINO team
  • custom OpenVINO operator authoring
  • direct interaction with OpenVINO team

Work Setup

  • Location: Bengaluru, India
  • Work mode: ONSITE
  • Employment type: Full-Time

Not Specified in JD

  • Salary range
  • Visa sponsorship
  • Remote eligibility
  • Travel
  • Security clearance
  • Coding test

What You'll Likely Work On

  • Deploy edge models on Intel NPU, dGPU, and iGPU while meeting defined SLAs
  • Maintain the OpenVINO build and quantization recipe, handling driver‑version compatibility
  • Optimize x86 CPU fallback paths using AVX‑512, AMX, and advanced threading strategies
  • Run and maintain the Intel device CI pool, detecting regressions across OpenVINO upgrades
  • Profile and tune CPU performance with VTune and perf, including assembly‑level hot‑loop analysis

Good Fit If You Have

  • Experience with AVX‑512 or AMX intrinsics is a strong plus
  • Prior direct interaction with the OpenVINO team or Intel ecosystem partners
  • Ability to author custom OpenVINO operators

Skills

  • OpenVINO
  • ONNX Runtime
  • Intel NPU / GPU
  • x86 CPU profiling & optimization
  • Model quantization
  • AVX‑512 / AMX intrinsics (plus)
  • VTune & perf
  • Custom OpenVINO operator authoring (plus)