Duncan Cameron-Steinke

Kelowna, British Columbia, Canada · (647) 936-0900 · dcameronsteinke@gmail.com

Quantitative researcher and engineer completing an MSc in Mathematics and Statistics, researching uncertainty-aware deep learning and regime-switching volatility forecasting in equity markets. Three years in industry to senior data engineer, plus two years co-founding an analytics startup.

Download PDF

Research Experience

Research paper (in preparation)

Belief Distribution Learning

March 2026 – Present

A novel deep learning algorithm with learned predictive uncertainty.

Python
TensorFlow
Keras
Machine Learning
Deep Learning
Uncertainty Quantification

Poster Slides

Derived a deep learning algorithm grounded in belief theory by combining Subjective Logic (Jøsang, 2016) with Evidential Deep Learning (Sensoy et al., 2018).
Introduced per-label Dirichlet modeling, estimating uncertainty independently for each label rather than as a single distribution over the full label simplex.
Evaluated against published Label Distribution Learning algorithms on standard benchmark datasets, demonstrating improved accuracy and better-calibrated uncertainty estimates.

Research project

Volatility Forecasting with Uncertainty-Aware Covariates

July 2025 – April 2026

Time series forecasting of hidden Markov model regimes with adaptive deep learning covariates.

Python
R
Hidden Markov Models
Time Series Forecasting
Deep Learning

Report Slides

Added forecasting to the hidden Markov model package hmmTMB.
Coupled an HMM with an uncertainty-aware deep learning model over covariates using a custom loss function.
Demonstrated improved forecasting of volatility regimes in the TSX composite index compared to HMM or deep learning alone.

MSc thesis, UBC Okanagan

Forecasting Radio Recombination Line Science

September 2025 – July 2026

Forecast and design of a next-generation radio telescope survey.

Python
Math
Statistics
Astronomy
Research

Interactive dashboard

Statistically analysed the radio environment to find radio quiet spectrum.
Computed the telescope’s sensitivity and expected science performance in detecting radio recombination lines.
Suggested smoothing and averaging strategies to improve the signal-to-noise ratio.
Optimized dish array geometry to improve signal resolution.

Experience

Research Associate

The University of British Columbia

Researcher with Dr. John Thompson on statistics, clustering, and deep learning methodologies.

September 2026 – January 2027

5 mos

Research
Machine Learning
Deep Learning
Python

Machine Learning Research Intern

Institute for Infocomm Research · Singapore

Authored a research paper (arXiv preprint) on short-text classification, developing a novel unsupervised text expansion method that generates new keywords directly from a pre-trained language model without requiring additional datasets or training.

Improved short-text classification accuracy under limited training data by generating 3–10 topically relevant words per target topic, validated against benchmark news headline datasets.

Research paper (arXiv)

Singapore International Pre-Graduate Award

January 2019 – April 2019

4 mos

NLP

Nano-Materials Research Intern

Sargeant Group · University of Toronto

Research internship in the field of electrochemistry studying catalyst nano-structuring.

April 2018 – August 2018

5 mos

Research

Senior Data Engineer

ServiceTitan

Embedded with executive, business development, and customer success teams to turn open business problems into shipped analytics, including a contract renewal dashboard that armed agents with product lift metrics ahead of renewal calls and a call center dashboard giving managers conversation and resolution statistics to prioritize reviews and evaluate agents.

Built and owned the business unit’s data platform from scratch on Snowflake and Tableau, managing one engineer and modeling billions of records into the metrics the unit ran on.

Designed the data model and owned the pipelines for a client-facing analytics dashboard serving 1,800+ customers (Snowflake, SQL), partnering with front-end engineers on the React implementation.

Designed the A/B testing methodology for a chat assistant prototype automating call center support, running sample-size calculations up front and significance testing on results to establish real effects, informing the product iterations that reduced agent handling time.

August 2022 – April 2024

1 yr 9 mos

Snowflake
Tableau
SQL
AWS
Production Software
Data Engineering
A/B Testing
Teamwork
Leadership

Analytics Engineer

Schedule Engine (acquired by ServiceTitan)

Established the company’s single source of truth for reporting, defining agreed-upon business metrics where none previously existed and building the data ingestion, cleansing, and augmentation pipelines to consolidate the underlying data.

Fully automated the client billing workflow, eliminating a recurring manual operations process.

April 2022 – August 2022

5 mos

Business Intelligence
Data Pipelines

Research Engineer II, Data Science

HomeX

Led development of a client analytics workflow that improved customer insights and outcomes, opening new revenue potential and deepening client relationships.

Deployed data ingestion microservices in Python supporting production analytics workloads.

Contributed to a marketplace ranking algorithm matching supply and demand across the platform.

July 2021 – February 2022

8 mos

Python
Analytics

Co-Founder & CTO

Backr.ai (YouTube analytics startup)

Co-founded a creator analytics platform for YouTube channels, building it full-stack in Python from real-time data ingestion through the customer-facing React application.

Raised $120K in non-dilutive funding and won first prize in the DDQIC pitch competition ($30K).

Selected for four accelerator programs: NextAI Montreal (2021), LaunchYU (Summer 2020), League of Innovators Labs (Batch 3), and QICSI (2019).

Hired and led a team of three full-time developers and a data analyst, owning product roadmap and technical strategy from inception.

Trained and deployed a custom NLP model for YouTube comment classification and sentiment analysis.

Architected a real-time streaming analytics platform on a microservice architecture hosted on Google Cloud.

Grew the platform to 50 creators, including multiple accounts with 3M+ subscribers.

June 2019 – July 2021

2 yrs 2 mos

Entrepreneurship
Leadership
Product Management
Full Stack
Python
SQL
NLP

Technology & Consulting Analyst (Co-op)

LoyaltyOne

Developed a new customer segmentation technique for marketing analytics by adapting statistical methods from sport analytics to a retail loyalty context.

Delivered analytics findings and actionable recommendations directly to the executive team, informing company strategy decisions.

September 2018 – December 2018

4 mos

Education

The University of British Columbia

Master of Science (MSc) · Kelowna, BC

Mathematics and Statistics

GPA: 4.0

Coursework in non-parametric models, stochastic processes, differential equations, generalized linear models, and optimization.

September 2024 – August 2026

Queen's University

Bachelor of Applied Science (BASc) · Kingston, ON

Engineering Physics

GPA: 3.9

Core studies in physics, mathematics, statistics, electronics, and computer programming.

Dean's Scholar
H. Jansen Memorial Scholarship (Awarded to the student with the highest standing in Physics I & II)

September 2015 – May 2020

Skills & Certifications

Skills & Technologies

A/B Testing
Analytics
Astronomy
AWS
Business Intelligence
Data Engineering
Data Pipelines
Deep Learning
Entrepreneurship
Full Stack
Hidden Markov Models
Keras
Leadership
Machine Learning
Math
NLP
Product Management
Production Software
Python
R
Research
Snowflake
SQL
Statistics
Tableau
Teamwork
TensorFlow
Time Series Forecasting
Uncertainty Quantification

Certifications

MIT Schwarzman College of Computing

Professional Certificate · Remote

Data Science and Machine Learning: Making Data-Driven Decisions

2022

Interests

Outside of my research, I enjoy canoe tripping and backcountry camping. My outdoor highlights include a 2-week trip by float plane to the Bloodvein River in Northern Manitoba, multiple trips to Algonquin Park and Temagami in Northern Ontario, Nanika-Kidprice in Northern BC, and a mountaineering trip to the Peruvian Andes.

After high school I deferred university for a year and spent 12 months living on a 39-foot sailboat, which my father and I navigated from Toronto to the Caribbean and across the Pacific to Australia. I love remote trips for their mental and physical challenges and the reward of accomplishing a goal with friends and family.

Duncan Cameron-Steinke in mountaineering gear on a snow-covered summit, with glaciated peaks and cloud below.

Duncan Cameron-Steinke

Research Experience

Belief Distribution Learning

Volatility Forecasting with Uncertainty-Aware Covariates

Forecasting Radio Recombination Line Science

Experience

Research ▾

Research Associate

Machine Learning Research Intern

Nano-Materials Research Intern

Industry ▾

Senior Data Engineer

Analytics Engineer

Research Engineer II, Data Science

Entrepreneurship ▾

Co-Founder & CTO

Co-op ▾

Technology & Consulting Analyst (Co-op)

Education

The University of British Columbia

Queen's University

Skills & Certifications

MIT Schwarzman College of Computing

Interests