I turn messy data science problems into production-grade software.
I have dual Master’s degrees in Applied Statistics and Computer Science, but what actually gets me going is messy data. I’m the kind of person who’ll happily spend hours digging through a broken dataset to find what’s hiding underneath.
Over the last few years I’ve built data pipelines, fine-tuned LLMs for production, designed dashboards, and written more validation scripts than I can count. I’m comfortable on the technical side and equally comfortable explaining what the numbers mean to people who don’t live in spreadsheets.
🛠️ Technical Arsenal
| Domain | Key Technologies |
|---|---|
| AI & LLM Engineering | Llama 3.2 (Fine-tuning), LoRA/UnslothAI, RAG Pipelines, spaCy, BERTopic |
| Cloud Architecture | AWS Serverless (Lambda, DynamoDB Streams, API Gateway), CloudFormation (IaC) |
| Backend & DevOps | Python (FastAPI/Django), Docker, GitHub Actions CI/CD, PostgreSQL |
| Data Engineering | ETL Pipelines, Web Scraping (Selenium), Visual Analytics (Dash/Plotly) |
🚀 Featured Engineering
01. AI-Powered Mindful Eating Companion
Generative AI & Mobile Engineering
The Challenge: Generic health apps lack context. I needed a system that could “read” user personality and adapt its advice dynamically.
- The Engineering: Fine-tuned two Meta Llama 3.2 (3B) models using LoRA adapters. Built a proprietary dataset of 1,500 expert-verified tips to ensure clinical safety.
- The Impact: Achieved ~88% relevance in user-rated pilot tests, running effectively on consumer-grade hardware.
02. SecureTask: Serverless Identity Verification
Cloud Architecture & Security
The Challenge: Building a secure biometric task platform without the operational overhead and cost of managing GPU servers.
- The Engineering: Architected a fully serverless AWS stack. Integrated Rekognition for face-ID (Selfie vs. ID) and used CloudFormation to deploy the infrastructure in <5 minutes.
- The Impact: Reduced infrastructure costs by ~40% compared to legacy containerized solutions.
03. Public Tenders Intelligence Dashboard
Data Engineering & Visual Analytics
The Challenge: Procurement analysts were losing hours manually digging through 10 years of unstructured government tender data.
- The Engineering: Built an automated ETL pipeline to clean data noise by 35% and applied BERTopic (NLP) to surface hidden spending trends automatically.
- The Impact: Accelerated trend identification speed by 30%, transforming raw CSVs into strategic intelligence.
🧩 Off the Clock
Outside work, I get absorbed in narrative games (Witcher 3, RDR2, Kingdom Come Deliverance…), psychological thrillers, and indie music. Strong black coffee required. Always…
Let’s Build Something Scalable
Current Status: Based in Halifax, NS. Open To: Roles in AI Engineering, Backend Development, Data Science, DevOps, Technical Support, and Data Engineering.