Laxmen Murali

Experience

• 

  Formula Manipal

  Feb 2020 - Mar 2022

  Senior Member Technical

  • Implemented Engine Control Unit Tuning and Battery Management System• Wiring harness / Electrical routing• Customized data logging for sensors at desired frequencies• Control of clutching and gear shifting system• Built CAN interface for easy data transfer to other controllers or data loggersKey Achievements:• Achieved a National rank of 2nd in EV and 3rd in CV Category at Formula Bharat Virtuals 2021-22• Achieved a National rank of 6th in EV and 8th in CV Category at Formula Bharat Virtuals 2020-21 Show less

• 

  MIT Gaming Industry Club

  Apr 2021 - Jul 2023

  Professional team member

  Run n Gun - Valorant eSports teamQualified for National level eSports competition amongst the universities of India by Elite Sports India.

• 

  HOPE Foundation

  Apr 2021 - Jun 2023

  Project Team Member

  • Created a survey to analyze the understanding of sustainable menstruation among young adults• Conducted webinars on sustainable menstruation for school and college students

• 

  ADIGOS

  Jun 2022 - Jul 2022

  Full Stack Developer
• 

  Samsung Semiconductor

  Jan 2023 - Jul 2023

  AI/ML Compiler Engineer

  • Built a high-performance computing system (supercomputer) based on RISC-V architecture, targeting Green500 rankings.• Achieved a 30% performance boost by leveraging compiler and loop optimization techniques, SIMD vectorization, and Intel VTuneprofiling, to enhance matrix multiplication—a core neural network operation—by identifying and resolving performance bottlenecks.• Implemented OpenMP for multi-threaded parallel processing and MPI for node-to-node communication in distributed systems.• Deployed Dockerized PyTorch and DL frameworks, emulating RISC-V CPU nodes via QEMU to train and infer models like ResNetand GPT, conducting layer-wise CPU time analysis to identify and resolve bottlenecks.• Created visual representations of neural network graphs for large language models such as GPT-2, ResNet-18, and ResNet-50 fromHuggingFace, using PyTorch Visualizers (PyTorchViz and TorchSummary).• Streamlined computational efficiency through model quantization, reducing precision requirements while maintaining accuracy.• Proficient in Intel VTune, PyTorch, TensorFlow, Docker, QEMU, OpenMP, MPI, GPU/CUDA, and familiar with LLVM/MLIR Show less

• 

  Samsung Semiconductor

  Apr 2024 - now

  Senior Software Engineer

  • Experience in working with Linux device drivers.• Led chipset 1580 and 1480 bring up for A & M series mobile phones, focusing on SD and eMMC storage.• Documented SD_MMC readiness for SoCs, ensuring alignment with development and verification standards.• Successfully implemented the SD support in bootloader.• Currently working on UFS, focusing on system initialization and optimization.