Systems and network administrator with an MSc in Internetworking. I keep networks routing, servers answering, and users working.
Edmonton, AB / open to remote roles across Canada$ traceroute mehedi.career
1habiganj-net.bdSystem AdministratorSep 2022 – Jun 202412 ms ↓ 2ualberta.caMSc Computing Science, InternetworkingCompleted 202618 ms ↓ 3bestbuy.caSolutions AdvisorJun 2025 – Mar 20269 ms ↓ 4geeksquad.bestbuy.caGeek Squad Agent, Technical ServiceMar 2026 – present4 ms ↓ 5* * *_Next hop: your team?? ms ↓I started out running production infrastructure at an internet service provider in Bangladesh, where I handled routing, DNS, and monitoring for a live network. I then came to Edmonton for a Master of Science in Computing Science (Internetworking) at the University of Alberta, with hands-on lab work across Cisco and Juniper gear, Active Directory, virtualization, and automation.
Today I work full-time on Best Buy's Geek Squad team, diagnosing and fixing hardware and software problems for customers every day. I'm looking for a Systems Administrator, Network Administrator, or IT Support role where I can put both sides of that experience to work.
Best Buy Canada, Edmonton, AB · Full-time
Best Buy Canada, Edmonton, AB
Habiganj Net Communications, Bangladesh
Buses in areas with patchy cellular coverage lose GPS, engine and ticketing data whenever they drop offline. I built an edge-to-cloud pipeline that treats outages as normal: each bus buffers every event locally and keeps seat sales in persistent storage so they survive power loss, then syncs to Azure IoT Hub with exponential backoff once the signal returns.
A Python cloud router sorts events into Azure Cosmos DB containers partitioned by vehicle. Every event records its sync delay, so operators can map cellular dead zones along a route, and passengers can look up any bus live on a Flask portal hosted on Azure App Service.
A computer-vision system that automatically finds missing fasteners and defective fishplates on railway tracks, two faults that can lead to derailments. I collected the images myself on the Akhaura–Sylhet line of Bangladesh Railway, labelled them, and trained and compared several object-detection models.
YOLOv5 Large and Extra-Large outperformed EfficientDet-D01 and Faster R-CNN (ResNet-152) on mAP@[0.5:0.95], and held up on track photos taken from angles the model never saw in training.
A scheduled AI agent that checks my inbox, Indeed, LinkedIn alerts, and the Alberta Job Bank each day, filters postings against target roles and locations, and logs new leads with notes on fit and legitimacy.
University of Alberta, Edmonton · 2026
Capstone: TransitSync offline-first fleet tracking
University of Liberal Arts Bangladesh · 2022 · CGPA 3.92
Capstone: deep-learning rail defect detection
Prefer the command line? Ask me anything with a few commands. Type help to start, or tap a button.