Highlights:

Skills:AI/ML, Python, C++, Data Science, Linux

Software Development Engineer 05/23/2022 - current
Advanced Micro Devices, Inc., Austin, TX United States
Industry: Semiconductor
I am working as a Software Development Engineer at AMD Research. My responsibilities and interest lie in exploring Machine Learning models and leveraging applicability to solve complex scientific problems, Performance Optimization, and Benchmarking.
Analyze LLM workloads with internal and external profiling tools: • Seek for bottlenecks for LLM inference on different generations of AMD hardware. • Scrutinize the bottlenecks and draw indications for future hardware/software modifications to achieve better performance (~117%) on AMD hardware. Inspect the applicability of AI-based surrogate models for PDEs without providing physics information: • Crafted an MLP model trained on 36 source locations (the frequencies of the source locations were 2 Hz and 5 Hz). The model was successfully able to generalize the pressure values on a 64x64 domain with an MSE of 0.001. • Accomplished the idea of piecewise model training in the time domain and submitted an IDF. Explore speculative decoding for accelerating LLM inference on AMD hardware: • Investigate the bottlenecks and performance of the draft model and target model pairs on AMD hardware. • Comparison of AMD hardware vs. other hardware. Investigate the bottlenecks and performance of the draft model and target model pairs on AMD hardware: • Authored validation scripts for software testing using Python. • CI/CD pipeline development using Jenkins and git-flow.--
Research Assistant 06/2021 - 05/22/2022
Texas State University - HiPE Lab, San Marcos, TX United States
Industry: HPC Research
I worked in this research lab on projects related to AI/ML, HPC, and performance analysis.
• Operated machine learning models for benchmarking and performance analysis on Dell AI workstations. • Compiled the ML models in Linux environments on Dell data science stacks and Nvidia docker containers. • Parallel processing of the ML models using multi-processes and multi-threads to increase efficiency. • GPU & Parallel Programming with C++, CUDA, OpenMP, and MPI to solve scientific problems. • Leveraging CPU/GPU Architecture to scale scientific problems in HPC environments.--

Khulna University of Engineering & Technology 04/2013 - 05/2018
Khulna, , Bangladesh
Degree: Bachelor's Degree
Major:Electronics & Communication Engineering
I started my journey toward engineering in my home country Bangladesh.

Login to view resume: Tamima_Rashid -