The RoleWe're hiring a
Senior Data Scientist to own machine-learning models
end to end - from framing the problem and designing the model through the training pipeline, deployment to a live endpoint and the
model-quality monitoring that keeps it accurate in production. This is a
hands-on, engineering-heavy data-science role: you build and ship your own models.
Where this role sits: you own the
models and their quality. Our
Data Engineering function owns the
shared data platform and serving infrastructure you build on - the warehouse, pipelines, governance, CI/CD and the operational reliability of the serving endpoints - so you stay focused on the science and the model lifecycle, not on running the platform.
We're explicit about this because it matters:
if you're looking for a pure research or notebook-only role, this isn't it. If you're energized by taking a model all the way to a live, monitored endpoint and seeing it drive real financial decisions - you'll thrive here.
We're
not building ML for ML's sake: we judge models by the
business outcomes they move - risk reduction, enrichment accuracy, customer adoption, operational efficiency, revenue and connecting your model improvements to those outcomes is part of the role.
What You'll Do- Own ML models end to end - frame the problem, write the design/RFC, build and train the model, ship it to a served endpoint and monitor its quality in production.
- Build your model's training pipeline and package it for serving, deploying onto the shared platform Data Engineering builds and operates.
- Own model quality, not the platform - you watch drift and performance and decide when to retrain.
- Evaluate rigorously - experimental design, statistical validation, drift detection and retraining, champion-challenger evaluation and promotion.
- Partner across the org - your model outputs feed the attributes/enrichment layer, payments risk, dashboards and client integrations, you'll collaborate with Data Engineering, backend, product and QA on contracts, deployment and rollout.
- Move fast with AI-assisted development -we use it to accelerate implementation and experimentation, the highest-leverage contribution in this role comes from strong problem framing, system design, evaluation rigor and clear technical specifications.
What You'll Work OnA few of the live ML systems this role owns end to end:
- Transaction categorization - a hierarchical, multi-task BERT classifier running as segment based models, trained on labelled data with synthetic top-ups for rare classes.
- Reversal / "final-category" detection - a hybrid ML + analytical-rule model classifying reversal types and linking each back to its original debit, promoted via champion-challenger.
- Transaction NER parser - a multilingual token-classifier extracting entity types from raw descriptions.
- Payments risk and balance forecasting - quantile time-series forecasting feeding a risk/offer decision layer.
- The enrichment suite - income / net-income, frequency detection, plus employment-loss, life-events and pay-frequency models.
- Model-quality and analytics tooling - the data-science team's own model-performance, PSI / drift and category-distribution monitoring for the models it owns.
Our stack- Python and SQL
- Google Cloud Platform
- BigQuery and modern data tooling
- PyTorch, HuggingFace and classical ML frameworks
- MLflow and Kubeflow
- FastAPI and containerized deployment
- Azure DevOps
You don't need experience with every tool listed above - strong production-ML fundamentals matter more than direct experience with our exact stack.
Python is the exception: it's a non-negotiable (see Key Requirements).
Why This Role- High ownership, low bureaucracy - you own models end to end and watch them drive real financial decisions for banks and fintechs.
- A modern, cloud-native ML stack with an AI-assisted development workflow.
- Real scale and real stakes - regulated financial data, enrichment and payments.
Key Requirements- Experience: 6-8 years building and shipping machine-learning models, including taking models to production yourself (training pipeline → served endpoint → monitoring).
- Education: Bachelor's degree in a quantitative field (Computer Science, Statistics, Applied Mathematics, or related), a Master's or PhD is an asset, not a requirement - production-ML ability matters more than credentials here.
- Non-negotiables: production-grade Python and the ability to take a model to a live, monitored service yourself, on a solid data-science / ML foundation. A notebook-only profile won't meet the bar.
- Work authorization: must be legally authorized to work in Canada.
Compensation RangeFor experienced and qualified hires located in Canada, of senior (IC4) level, the compensation range is between $120,000 to $160,000 CAD annually.
As part of the total rewards package, Flinks offers:
- Health & Dental coverage as of Day 1
- Flexible Paid Time Off (FTO)
- Remote work environment with frequent in-person gatherings and activities.
- Career development, learning opportunities and growth
- And more
