Summary
Senior Developer and Data Engineer with 6+ years of experience designing, building, and running
production data pipelines, microservices, and automation systems. Sole technical owner of an enterprise
data platform, responsible for architecture, development, deployment, stakeholder engagement, and
technical documentation across 30+ production systems. Results include $850K+ in revenue leakage
prevented and 19+ FTE hours saved daily. Primarily working in Python, SQL, Snowflake, FastAPI, Docker,
Kubernetes, and applied AI/ML. Background in customs brokerage and logistics operations.
Technical Skills
Languages
Python, SQL
Data
Snowflake, SQL Server, Redis, Pandas, ETL pipeline design
AI / ML
OpenAI API, LangChain, RAG pipelines, ChromaDB, NER ensembles, vision-language models
Infrastructure
Kubernetes (AKS), Docker, Azure (Key Vault, Blob Storage), KEDA, Git/GitHub
Frameworks
FastAPI, SQLAlchemy, Pydantic, spaCy, Transformers (HuggingFace)
Professional Experience
Remote / Hamilton, ON
- Designed and built a customs entry automation pipeline with AI tariff
classification (RAG + LLM). Multi-source ingestion across 100+ importer accounts, sub-15-minute
processing. 9.0 FTE annualized savings.
- Built ML and AI microservices in production: a multi-signal classification
ensemble (94.97% accuracy on 10,000+ ground truth data points) and a
vision-language model extracting 90+ fields from commercial documents.
- Created an ETL framework powering 30+ automated workflows, integrating cloud
data warehouse, relational databases, SFTP, and legacy mainframe sources.
- Automated high-volume operational workflows: shipment assignment (saving
19+ FTE hours daily), document uploads (485K+ processed), and refund claims.
Prevented $850K+ in revenue leakage.
- Led migration of all production services to Azure Kubernetes Service (AKS),
managing 20+ deployments and supporting infrastructure.
- Built internal web applications (FastAPI, JWT auth) including admin dashboards and
customer-facing portals.
- Achieved maximum performance ratings every year from 2022–2025.
Manager-recommended for promotion to Technical Program Manager.
Remote / Brampton, ON
- Reduced automated shipment profiling failures by 13% by auditing rule logic and
resolving system issues across the customs automation platform.
- Built work distribution and exchange rate maintenance tools, eliminating manual
tasks and saving 12+ FTE hours per day while reducing compliance incidents.
- Maintained and optimized datasets, SQL queries, and internal data models for customs clearance
operations.
- Developed automated workflows for duty classification, document processing, and mainframe data
integration.
- Provided system training to internal users and external customers on customs automation tools.
Remote / Brampton, ON
- Evaluated shipments for tax and duty using harmonized tariff codes, valuation methods, and trade
agreements.
- Handled billing, credits, and customer disputes with high accuracy and service quality.
- Supported development of a write-off reduction project later adopted as a
permanent workflow, saving over $2 million in its first year.
Hamilton, ON
- Coordinated preparation and shipment of construction materials in a distribution facility. Order
fulfillment, forklift operation, and inventory management.
Selected Projects
Customs Entry Automation
- Full pipeline: multi-source ingestion, AI tariff classification (RAG + LLM),
automated field transformations, distributed tracing, and multi-destination output. 20,000+ entries
generated; 100+ importers onboarded.
Entity Classification Service
- Multi-signal weighted ML ensemble for organization vs. person classification.
94.97% accuracy on 10,000+ ground truth samples. FastAPI microservice on Kubernetes.
Distributed Manifest Processing
- Distributed ETL with Redis queue, KEDA auto-scaling (0–20 replicas), worker
heartbeat recovery, and cloud data warehouse staging. Processes facility data across 20+ service
centers.
Vision-Language Document Extraction
- VLM service extracting 90+ structured fields from commercial documents (PDF/TIFF)
using GPT-4.1 vision with adaptive resolution and multi-page chunking.
Legacy Mainframe Integration
- Bridge between legacy mainframe and modern data infrastructure. 30+ automated
report extractors delivering to cloud data warehouse, relational database, and file-based
formats.
Education & Certifications
Professional Program in Data Science
Open University of the Netherlands (DHL Incompany Program) | 2022 – 2023
| 280 hours | Grade: 8.3/10
Regression, clustering, ensemble methods (XGBoost, Random
Forest), NLP, neural networks, PCA, cross-validation, Python/Jupyter.
Diploma in International Trade
Sir Sandford Fleming College, Peterborough, ON | 2013 – 2016
Certified Customs Specialist (CCS)
Canadian Society of Customs Brokers | 2019 – 2025