Mashal Rane

About Mashal Rane

12+ years experienced engineer with a demonstrated history of working in Semiconductor, Automotive in-vehicle infotainment and Automated Driving domain, highly skilled in developing code, technical designs, managing projects/programs, and teams.• Building and scaling software (ML and AI based) and hardware (NPU, DSP, GPU and CPU) accelerator deliverables in a highly competitive time-to-market environment for mobile and automotive domain.• Demonstrated capability as \'Subject Matter Expert\' for tuner team at Harman International, technically leading the team throughout program releases, planning program increments for agile release train, cross-domain interactions, facilitating scrum events, project estimation, planning, and risk management activities. • Armed with the skills of effective communication, time management, critical thinking, and problem-solving.

Experience

1. Senior Staff Engineer

   Samsung Semiconductor

   Mar 2023 — Present · Bengaluru, IN

   Managing a team of 4 which specializes in Graph processing and optimizations which best suites Exynos backend.• Led cross-functional teams to enable Geekbench AI on Samsung Exynos backends for Exynos 2400 and 2500, driving backend development, integration, validation, and delivery; enabled benchmark across 2 flagship SoC generations.• Core contributor and maintainer of Exynos AI LiteCore (Samsung AI Lightweight SDK), enabling efficient neural network inference across NPU, DSP, CPU, and GPU, and supporting external customers with stable, high-performance SDK releases.• Enabled graph compilation and execution for major ML frameworks including TensorFlow, PyTorch, and Caffe on Exynos hardware backends.• Hands-on experience using Google LiteRT Next and ONNX Runtime Execution Provider• Implemented and optimized graph-level performance optimizations for computer vision workloads, improving hardware utilization and inference efficiency on Exynos backends.• Improved ML model accuracy issues using Signal-to-Noise Ratio (SNR) analysis against reference outputs.• Contributed to industry-standard AI benchmarks such as AI Benchmark and MLPerf, strengthening Exynos competitiveness and visibility in performance evaluations.• Contributed to Samsung Exynos AI Studio used for model compilation and deployment.• Owned program management for AI software across the SoC lifecycle, covering requirements definition, pre-silicon planning, post-silicon bring-up, software development, verification, optimization, deployment, and maintenance.• Delivered on-device and off-device ML solutions for end-user applications• Contributed towards enablement of LLMs and generative AI models on Exynos backends, addressing compilation, performance, and hardware execution challenges for next-gen AI workloads.

Education

• Visvesvaraya Technological University

  B.E, Electronics and Communication

• Goa Board of Secondary and Higher Secondary Education (GBSHSE)

  High School, science

• Goa Board of Secondary and Higher Secondary Education (GBSHSE)

  Higher Secondary School, Science

  2007 — 2009

Skills

• Mysql
• Testing
• Engineering
• Embedded Systems
• Tortoise Svn
• C++
• Unix
• Autocad
• C
• Microsoft Powerpoint
• Software Development
• Eclipse
• Microsoft Office
• Matlab
• Linux