m
miran_29

Miran

@miran_29

Machine Learning Engineer

États-Unis
Anglais, Bengali
Certaines informations sont présentées en anglais.
À propos de moi
I’m a computer engineer graduate from Southern Illinois University Edwardsville. I have contributed towards developing CVIPToolbox for MATLAB. CVIPToolbox is a GUI developed for MATLAB environment to perform image analysis and processing tasks. I also worked as an intern at AMD as a Machine Learning Performance Software Engineer Intern. I contributed to the MIGraphX team by developing Python script to run performance on ML models and ONNX operators using different GPU to find performance gap.... Plus d’infos

Compétences

m
miran_29
Miran
hors ligne • 
Temps de réponse moyen de 1 heure

Voir mes services

Vision par ordinateur
I will build ai computer vision models using pytorch
Rédaction de script
I will create a custom python automation script for your task

Expérience professionnelle

Interco

System Administrator

Interco • Temps plein

Nov 2024 - Present1 yr 9 mos

Managed end-to-end Instantly.ai email campaigns, including lead data cleaning, segmentation, campaign setup, and importing leads into the CRM. Managed and resolved hardware, software, network, and user-access issues while monitoring and optimizing IT infrastructure performance. Administered Microsoft Dynamics 365 and FreePBX, supporting workflow automation and VoIP communications. Managed Active Directory, including user accounts, permissions, onboarding, offboarding, and access control. Troubleshot and maintained Windows systems, network devices, printers, and other IT hardware to ensure reliable business operations. Managed Lumi Center VMS and security camera systems, including system configuration, user access, troubleshooting, and maintenance.

AMD

Intern

AMD • Temps plein

Sep 2023 - Dec 20233 mos

Developed ONNX subgraphs from ResNet50, BERT, LLaMA, and Stable Diffusion models to benchmark ML operator and inference performance across approximately 15 operators. Built Python automation tools to execute inference workloads, collect performance metrics, and compare model execution across AMD and NVIDIA GPUs using MIGraphX, TensorRT, and AWS EC2 environments, identifying that NVIDIA TensorRT consistently outperformed AMD solutions across all benchmarked scenarios. Automated performance analysis and reporting workflows in Python, generating insights and presenting benchmarking results and technical findings to the MIGraphX engineering team to support optimization efforts.