Hi, I'm

Eric Backman

Data Scientist & Applied AI Specialist

I build data pipelines and AI systems that simplify menial tasks and reduce day to day overhead.

BMO · Toronto · Data Science · Python · Java · Data Engineering · Data Analytics · Queen's B.A.Sc.

About

I'm a data scientist and engineer based in Toronto, working at the intersection of AI systems and financial risk infrastructure. I am designing and implementing Agentic AI tools that automate and accelerate regulatory capital analysis. Agentic tools that let analysts do in minutes what used to take days.

My path started in the Canadian smart home space at Ecobee, where I built experimentation platforms and data pipelines for smart home devices. I've since moved through healthcare APIs, sensor validation, and ML model development before landing in financial risk.

>_

Agentic AI

Multi-agent orchestration systems that integrate LLMs into analyst workflows

$

Risk Domain

FRTB SA (Fundamental Review of the Trading Book) — regulatory capital expertise that most engineers lack

//

Full Stack

End-to-end delivery: data ingestion to API to analytics

Experience

Senior Analyst, AI Implementation

BMO
Aug 2025 — Present
  • Leverage data science experience to build agentic AI platform. Automates the extraction, comparison, and analysis of financial metrics across multiple asset classes — reducing investigation time from days to minutes.
  • Designed AI-assisted analytical tools that help analysts quickly identify trends and movements in large datasets.
  • Added monitoring and anomaly detection for internal data systems, catching data quality issues before they hit analysts.
  • Developed evaluation protocols with human review loops, documenting validation results and bias analyses for each module.

Data Scientist

Ecobee
May 2021 — Jul 2022
  • Validated air quality sensors for next-gen smart thermostat using field trial data.
  • Researched Model Predictive Control for thermal controller; managed testing infrastructure.
  • Improved smart security model accuracy ~6% using Bayesian inference and retrained models.

Junior Backend Developer

Smile CDR
Nov 2020 — May 2021
  • Built FHIR-compliant data mapping APIs in Java/Spring Boot for healthcare monitoring data.
  • Enabled asynchronous refreshing of third-party healthcare device data.

Data Science / Engineering Intern

Ecobee
Dec 2018 — Jul 2019
  • Built experimentation platform enabling C++ A/B tests on thermostat devices in the field.
  • Designed data pipelines (Apache Beam) to clean and organize experimental data.
  • Identified blocking bugs causing multi-month delays; built remote testing infrastructure.

Embedded QA, Test Developer Intern

Ecobee
May 2018 — Dec 2018
  • Automated integration testing and Alexa voice detection accuracy evaluation.
  • Became primary point of contact for Alexa-related issues across the company.
  • Built audio isolation fixtures for testing in an open floor plan environment.

Skills

Agentic AI

  • GitHub Copilot — Financial risk multi-agent system
  • M365 Copilot — Team chatbots and prompts
  • Claude Desktop — Personal projects, data analysis

Languages

  • Python
  • SQL
  • Java
  • Bash

Data

  • BigQuery
  • PostgreSQL
  • MySQL
  • SQLite

Infrastructure

  • Apache Beam
  • Dataflow
  • Airflow
  • GCP
  • Docker
  • CircleCI

Projects

NBA Data Explorer

Personal Claude-powered repository to answer ad hoc basketball questions I have watching games with my family.

Python · Claude · pandas · nba_api GitHub ↗

MTG Deckbuilder

Scanned my collection and uploaded as CSV, allowing Claude to assist in deckbuilding decisions and strategy.

Python · Claude · pandas · CSV GitHub ↗

Risk Knowledge Base (FRM Level 1)

A personalized tutor agent to prepare me to write the FRM Level 1. Built to allow Claude to give me endless practice exam questions and enable detailed follow-up questions.

Claude · Finance · Risk · FRM

Interactive Dive Map

World map of 150+ scuba dive sites across 12 countries. Click pins to see dive details, photos, and underwater video.

Leaflet.js · JavaScript · GitHub Pages Live Map ↗

Family Travel Map

Interactive map of 65 family trips across 42 countries — multi-traveler routes, clustered pins, and per-person travel stats.

Leaflet.js · JavaScript · Cloudflare Pages Live Map ↗

Job Hunting GitHub

Had Claude ingest my old resumes and descriptions of current work, used to create new resumes and this website.

Claude · HTML · CSS · Netlify GitHub ↗

Education

Queen's University

2015 — 2020

Bachelor of Applied Science — Applied Mathematics & Computer Engineering

Honours 3.77 GPA 16-mo Internship French Immersion