Open to opportunities

Matt Nelson

@mattnelson

Message

Data & analytics leader building AI-native data platforms and shipping LLM-in-the-loop products.

Canada

Message

What I'm looking for

I’m looking to lead an AI-native data/analytics team, architect production platforms end-to-end, and ship LLM-in-the-loop pipelines that scale. I want autonomy with modern tooling (Dagster/Snowflake/Polars/Terraform) and a high bar for correctness.

I’m a data and analytics leader with 9+ years of experience, architecting modern data platforms and shipping AI-native products. I build production-grade systems using Dagster, Snowflake, Polars, DuckDB, and dbt—then layer in LLM-in-the-loop pipelines and structured-output inference to replace brittle rules.

At StackDX, I led the USA data platform rebuild, replacing a legacy Python ETL stack with a production-grade Dagster-native architecture on AWS. I’ve operated at scale (1.4B+ mart rows, 400M+ downstream exposure rows) and cut critical-dataset processing time by up to 98% using targeted Polars and DuckDB rewrites, while standardizing 100+ assets with reusable patterns.

As founder and principal engineer at twochannel.ai, I designed an AI-native monorepo that maintains 16k+ products across 2k+ brands with minimal manual work. I use an end-to-end Claude Code toolchain (skills, sub-agents, hooks, MCP servers) to run LLM-augmented ingestion and human-reviewed semantic dedup, and I ship the catalog and search experience with Next.js, Directus, and Algolia.

Earlier, I built Canlin’s industry-leading analytics foundation on dbt Cloud and Snowflake, integrated SCADA and operational systems into a single source of truth, and owned advanced analytics and ML Ops. I also led Canlin’s annual NI 51-101 corporate reserves for 5 years—technical revisions added +$730M NPV10 over 4 years—while mentoring data professionals and advocating for sustained investment in modern, cloud-first analytics.

Experience

Work history, roles, and key accomplishments

Current

Founder & Principal Engineer

Current

Twochannel.ai

Founded and built an AI-native monorepo powering an end-to-end product catalog, ingestion, search, and recommendations. Maintained 16k+ products across 2k+ brands using LLM-in-the-loop pipelines, structured-output enrichment, semantic dedup, and a Claude Code toolchain, then shipped a Next.js/Directus/Algolia/Neo4j platform with CI-grade E2E testing.

Claude Code MCP Servers Structured Outputs Next.js Directus Algolia Neo4j TurboRepo

Current

Data Platform Engineer

Current

Stack Technologies Ltd.

Architected a production Dagster-native data platform on AWS, transforming 1.4B+ mart rows and 400M+ downstream exposure rows. Cut critical-dataset processing time by up to 98% with Polars/DuckDB rewrites, and led Terraform IaC plus CI/CD; built Claude Code tooling to accelerate pipeline shipping across 15+ US states and 400+ transformation assets.

AWS Terraform Polars DuckDB DBT CI CD GitHub Actions Claude Code Dagster

Data & Advanced Analytics Lead

Canlin Energy Corporation

Designed and ran Canlin’s dbt Cloud + Snowflake data stack integrating SCADA, production, reserves, financial, and accounting systems into a single operational analytics foundation. Built an XGBoost 30-day production forecaster for 8,000+ wells, deployed anomaly/portfolio risk models and ML ops workflows, and delivered automated weekly production reporting that saved 50+ hours/week across Operatio

Dbt Cloud Snowflake DataRobot Tableau Streamlit XGBoost Snowflake Streams Tasks Snowpark SCADA Integration

Senior Data Analytics Engineer

Paramount Resources Limited

Delivered modern BI and data-engineering solutions for corporate reporting, including multi-level netback and operating cost dashboards in Power BI and automated weekly well production reporting (Spotfire + Power Automate). Built Streamlit web apps to replace manual workflows with optimized scheduling/forecasting logic and led a local modern data-stack PoC using Python/dbt/cube.js/DuckDB/PowerBI i

Power BI Spotfire DBT DuckDB SQL Python Streamlit Data Visualization

Corporate Data Scientist & Reserves Manager

Canlin Energy Corporation

Owned Canlin’s annual NI 51-101 corporate reserves process and automated the reserves data lifecycle (generate → clean → compare → visualize → report → share) to improve accuracy and support auditors and evaluators. Led Tableau implementation for company-wide self-service reporting and produced 100+ data sources, 50 workbooks, and 20 Prep flows; technical revisions added $730M NPV10 over 4 years.

Tableau Data Warehousing NI 51 101 Reserves Data Governance NPV10 Analysis Data Validation Self Service Analytics Petroleum Data Expertise