Staff Data Scientist - Growth

Canva Canva · Enterprise · Sydney, Australia · Information Technology

Staff Data Scientist focused on Growth, responsible for designing and building causal models to quantify relationships between input metrics and foundational goals (paid upgrades, user retention, acquisition). The role involves partnering with cross-functional teams to design experiments, validate causal inference, and embed findings into the Semantic Layer. Additionally, the role will act as an analytical lead for Monthly Business Reviews, drive accountability for output metrics, and serve as a key analytical partner to leadership. A key aspect is developing AI-powered tools and agents to scale analytical capabilities.

What you'd actually do

  1. Design and build causal models that quantify the relationships between input metrics (feature exposure, adoption, engagement, etc.) and component metrics to Canva's foundational goals (paid upgrades, user retention, user acquisition, etc.) — establishing which hypothesised links hold up and how strong they are.
  2. Partner with embedded data science and analytics teams to design experiments and/or analytical approaches that test causal hypotheses, adapting methodology to the wide variety of contexts across Canva (short-run A/B tests, long-term holdouts, quasi-experimental methods where randomisation isn't feasible).
  3. Own the causal layer of Canva's Metrics Tree — validating and quantifying the links between team-level input metrics and company-level outcomes, and feeding findings back into how teams set targets and define success. Over time, extend this framework to cover the UCM funnel as it matures.
  4. Act as the analytical lead for Monthly Business Reviews in partnership with FGP Ops — onboarding team metrics, sense-checking what's being tracked, owning the MBR frontend in Airtable, and ensuring the MBR surfaces the right signals to leadership.
  5. Partner with Analytics Engineering to embed validated metrics and causal findings into the Semantic Layer — enabling dimensional cuts, and the tooling teams need to track and investigate key metrics.

Skills

Required

  • Causal inference
  • Econometric methods
  • Instrumental variables
  • Difference-in-differences
  • Synthetic control
  • Structural causal models
  • Experimental design
  • SQL
  • Python
  • Production-quality analytical code development
  • Strategic communication
  • Commercial instinct
  • Forming and defending a point of view
  • AI and automation for analytical capability

Nice to have

  • feature stores
  • ML data engineering
  • fine-tuning
  • model serving
  • agent orchestration
  • tool use
  • evals
  • guardrails
  • llm_observability
  • rag
  • vector_db
  • frontier_research
  • interpretability
  • synthetic_data
  • agent_research
  • rl_post_training
  • rlhf
  • reward_modeling
  • rl_robotics
  • embodied_ai

What the JD emphasized

  • deep expertise in causal inference and econometric methods
  • strong experimental design skills
  • fluent in SQL and Python as daily working tools
  • build production-quality analytical code
  • transition between deep technical work and strategic communication
  • strong commercial instinct
  • comfortable forming and defending a point of view
  • Develop AI-powered tools and agents that amplify your analytical reach
  • curious about how AI and automation can extend analytical capability