2021 - 2025Canada - Montréal,hybrid
IoT & Smart Building Platform Engineer @Bell
Bell is a very well know historic telecommunication company in Canada. I've joined the team to help
them develop their smart-building plateform. Later my role in the team evolve as I was then charged to design
complex feature as well as implement them. The trusting relationship we built at the time helped the team to
be one of the most successful among the all the IoT teams @Bell.
Phase 1: First steps in the team
IoT Data Ingestion & Real-time Monitoring Developed serverless data processing pipelines to ingest and analyze telemetry from industrial diesel generators.
My achievements:
- Built a reporting engine to automate generator activity tracking (runs and test cycles), providing critical operational insights to stakeholders.
- Improved database resource utilization
- Successful onboarding on UI, the team was able to provide real time alarms page
Key metrics
- Trunk base versioning allowed us to deploy multiple times during the week
- The first weeks the application ingested ~2To of telemetry data (devices were spamming the plateform because of a configuration issue) (bad data is still data )
- Achieved a 66% reduction in database resource cost
Stack:
Node.jsJavaAzure FunctionsAzure CosmosDBTerraformKubernetesDockerVue.js
Phase 2: Lead on new features design & implementation
Enterprise Platform Migration & Scalability Supported the strategic migration of vertical IoT solutions into a unified, large-scale enterprise platform.
My achievements:
- Rebuilt the reporting engine from the ground up to meet new high-scale customer requirements and deployed the core infrastructure for the migrated services.
- Specified and built an aggregator service to enable the platform to process telemetries from related assets
- Identified and solved a race condition when the platform was ingesting and processing incomming messages
- The plateform became device agnostic
Key metrics
- Feature with high interest where added to the platform enabling more customer use case
- Proposed planning cut previous estimation in half thanks to a better architecture
- Noise caused by the race condition completely disapeared
Stack:
Node.jsTypescriptVue.jsJavaSpring BootQuarkusKubernetes (AKS)AzureTerraformRedisCosmosDBMongoDBGitLab CI/CD
Phase 3: Data engineering
Real-Time Data Engineering & DevOps Excellence. Designed a new spark pipelien to ingest real time genrator telemetries and calculate fuel consumption. Tasked with re-engineering a legacy real-time geolocation tracking pipeline to improve reliability and maintainability. Plus integration of new kind of assets with grouping / splitting telemetries capabilities.
My achievements:
- Successfully migrated a PySpark/Azure Synapse pipeline to a modern development workflow. Introduced unit testing, Dev Containers for environment parity, and fully automated GitLab CI/CD pipelines to ensure stable production deployments
- First version of the testing utility tool for notebook allowing us to run automatic tests during GitLab CI/CD pipeline
- Thanks to my design proposition one service could be improved instead of going through a full rewrite
Key Metrics
- Smoother deployment of the new pyspark pipeline into production. Reduced release cycle time from months to weeks
- The entire team was able to start contributing on the project thanks to the devcontainer
- Testing utility tool developped was then used by other teams
Stack:
PythonPySparkAzure SynapseGitLab CI/CDDev Containers
Cross-Functional Contributions:
- Quality Assurance: Enhanced the testing ecosystem using Robot Framework to develop E2E test suites, creating custom keywords for Azure Event Hub to validate asynchronous event-driven architectures.
- Devx : improve Gitlab pipeline by making it quicker and also provide new features to company own tools
- Agile Collaboration: Actively participated in Scrum ceremonies, contributing to architectural design and technical strategy.