Soumik Farhan Email & Phone Number

Edmonton, AB, CA

Edmonton, Alberta, Canada

• Currently working on LLM agent for building up automatic task processing for large workforce companies.
• Explored key heuristics in shift scheduling data for large workforce companies using open large language models (LLMs) such as Llama, and Mixtral; functional callings api’s for other LLMs such as GPT 4o, and Jamba.
• Researched and developed a custom ReAct-based LLM agent framework for extracting information flow.
• Presented SOTA papers relevant to current trends in LLM to technical and non-technical audiences

Toronto, Ontario, Canada

• Integrated ML-based image segmentation pipeline for Intracranial Volume Extraction usingdocker and cloud computing for ADNI cohort’s (DICOM and nifti) dataset.
• Designed and developed brain imaging data format to standardized BIDS format. This processreduced human dependence and integrated subsequent analysis with existing neuroimaginganalysis tools reducing the annotation.
• Presented 20+ SOTA papers as posters with reproduced results. Proposed a research roadmapleading to multiple unique applications of Graph Convolutional Networks across multiple modalitiesof neuroimaging data.
• Investigated Brain Connectivity Analysis for predicting neurodegenerative diseases in advanceusing Graph Attention Network.
• Assembled the data processing pipeline for dMRI, fMRI, image registration, atlas generation, andvolumetric & cortical thickness analysis using containerized environments (docker, singularity).
• Performed code review via git, run/test/benchmark/deploy the solutions on GCP/Compute Canada.

Edmonton, Alberta, Canada

1. Worked as a Machine Learning Intern at AMII on the Canada Wildfire Project for predicting extreme wildfires 3-5 days in advance for any specific coordinate.2. Curated a balanced dataset of Western Canada’s (approximately 1M sample) synoptic scale weather variables for advanced wildfire prediction for a specific geolocation.3. One of the unique aspects.

Winnipeg, Manitoba, Canada

1. Conducted research and development in areas of brain connectivity, computationalneuroscience, deep learning, and neuroimaging data (fMRI).2. Implemented Graph CNN and U-Net-based structures for mapping brain structure to functional connectivity.3. Experimented with different architectural combinations of Graph CNN and U-Net for optimizing the model.

Master Of Science - Ms, Biomedical/Medical Engineering

Completed Master of Science from Biomedical Engineering Graduate Program with a thesis on " Predicting subject‐specific brain functional.

Master Of Science (M.Sc.), Biomedical/Medical Engineering

1. Completed Master of Science in Biomedical Engineering with a specific focus on biomedical signal processing. 2. Completed thesis on.

B.Sc In Electrical And Electronics Engineering, Electrical And Electronics Engineering, Biomedical Engineering, Telecommunication

Activities and Societies: IUT IEEE STUDENT BRANCH IUT Photographic Society IUT Islamic study Society An active member of the IEEE student.