Data Scientist 2

Data Scientist - 2 (DS-2)

Job family: Data Science
Level: DS-2 (equivalent to MLE-2 / AIE-2)
Scope of impact: Feature
Theme: Grows and Acts completes scoped modelling tasks and
improves team process

Why this role exists

Product outcomes need deep problem ownership and tight iteration
with PMs and product engineering. A DS-2 turns a scoped product
problem into a calibrated model or decision system ships it
through the standard production path and owns its performance
after launch. You operate with minimal guidance on a defined
feature not the whole domain.

What you own

- A scoped modelling problem framed as a DS task: hypothesis
success metric offline and online evaluation plan.
- Calibrated predictive or causal models with well-behaved
probabilities and effect estimates.
- Repeatable pipelines integrated with production workflows not
one-off notebooks.
- Basic model monitoring for the features you ship.
- Post-launch performance of your model and its link to the
target KPI; iterate using telemetry.

What you do not own (yet)

- Platform uptime and shared serving infrastructure (ML
Engineering owns this).
- Domain-wide priority setting across multiple initiatives (DS-3
and above).

What youll do (proficiency expectations at L2)

Data-driven decision making
- Build calibrated predictive or causal models with sound
probability and effect estimates.
- Articulate the impact of uncertainty and select an applicable
course of action with minimal guidance.
- Stress-test findings with simple mental models or simulations
before trusting them.

Technical expertise
- Set up fully reproducible environments for your own work and
share the guides with peers.
- Package work into repeatable pipelines and integrate them with
production workflows.
- Implement basic model monitoring.

Applied ML/AI/DS
- Frame and scope an opportunity as a DS problem and pick the
right solution family (prediction optimization causal).
- Review recent literature build reproducible pipelines and
fairly compare alternative models.
- Run controlled pilots that connect model uplift to a target
KPI.

Experimentation and inference
- Frame a testable hypothesis and pick the right design (A/B or
hold-out).
- Run multi-metric or stratified tests with power checks and
CUPED variance reduction.
- Conclude using confidence intervals state the limitations
and tie results back to a target KPI.

Strategy and influence
- Scope an opportunity into a well-posed DS problem naming the
RoI and the product and process changes it implies.
- Align stakeholders on the KPI leverage of a proposed approach
and secure agreement on scope and goals.
- Coordinate with engineering and product leads to launch
features where the model provides core value; shape planning and
risk assessment.

How you work with others

- PM: co-own the outcome and prioritisation for your feature.
- Product Engineering: integrate your model into customer-facing
experiences.
- ML Engineering / AI Engineering: consume platform primitives;
collaborate on evaluation reliability gates and production
readiness.

What we expect from a strong DS-2

- Ships production artifacts on the standard path not
prototypes that stall at the production boundary.
- Improves at least one team process (templates reviews
reproducibility) beyond their own tasks.
- Owns outcome integrity: model outcomes stay aligned with
product outcomes after launch.