Graduate Data Engineer / Analyst

Job description · fictional employer

Graduate Data Engineer / Analyst

Yarrow Energy Retail

Location: Brisbane - hybrid, 2 days in office
Employment type: Full-time, permanent - 12-month graduate program
Salary: AUD $80,000 base + 12% superannuation
Reports to: Data Platform Lead
Intake: February 2027 - applications close 10 October 2026

About us

Yarrow Energy Retail sells electricity and gas to about 210,000 households and small businesses across Queensland and New South Wales. Every one of them generates meter reads, billing events, payments and service calls. Our data team turns that into pricing decisions, hardship identification, regulatory reporting and the forecasts the trading desk relies on.

The team you would join

The data group is twelve people: five data engineers, four analysts, two analytics engineers and a lead. You would split the first year across engineering and analytics rather than choosing on day one, because the best people in this field can do both and most graduates do not yet know which they prefer.

What you will do

Build and maintain data pipelines in Python and SQL, orchestrated in Airflow
Write dbt models with tests and documentation - an untested model does not get merged
Investigate data quality issues end to end, from a stakeholder saying the number looks wrong to a fix in the source
Build and maintain Power BI reports that people actually open, and retire the ones they do not
Work directly with billing, hardship and trading teams to understand what a number is for before you produce it
Support regulatory reporting cycles, where accuracy and traceability are non-negotiable
Contribute to the data dictionary and lineage documentation
Present findings to non-technical stakeholders, including saying when the data cannot answer the question

What we are looking for

A completed or in-progress bachelor degree in data science, computer science, IT, mathematics, statistics, engineering, economics or a related discipline
SQL you can defend in an interview - joins, group by, having, window functions or a demonstrated willingness to learn them fast
Python for data work: pandas or equivalent, reading awkward files, basic scripting
The habit of questioning a result that looks too good
Clear communication in writing. Half of analytics is explaining a number to someone who did not ask for the caveats
Attention to detail - in energy retail, a wrong number becomes a wrong bill
Full Australian working rights

Nice to have

Exposure to a cloud data warehouse - Snowflake, BigQuery, Redshift or Databricks
dbt, Airflow or any orchestration tool
Power BI, Tableau or Looker
Statistics beyond an introductory unit - regression, hypothesis testing, time series
Any exposure to the energy sector, or to a regulated industry
Version control for analysis work, not just for software

Our stack

SnowflakedbtApache AirflowPython (pandas, Polars)SQLPower BIAWS S3FivetranGreat ExpectationsGit

What the program gives you

Rotation across data engineering and analytics in the first year, then you choose
A named business stakeholder from month two, so you learn the domain and not just the tables
dbt and Snowflake certification paid for, with study time
$2,000 learning budget
A team that writes tests for data and treats analysis code like code

How the process runs

1
Application
CV plus a link to any analysis you have done, in any format.
2
SQL screen
45 minutes, live but collaborative. Realistic messy tables, and we help if you get stuck.
3
Case exercise
Take-home, around three hours. A dataset with deliberate problems in it, and a business question.
4
Case discussion
60 minutes - present your findings to two people, one of whom is not technical.
5
Team interview
45 minutes - stakeholder management, judgement, how you handle being asked for a number you do not trust.
6
Offer
Within a week of the final stage.

The best graduate analysts we have hired were not the ones with the most tooling on their CV. They were the ones who asked why a number was being requested before producing it. Reasonable adjustments are available at any stage.

Example CV · fictional candidate

Ella Marchetti

Written to the job description on the previous tab. Notes on the right explain each choice.

Ella Marchetti

Graduate Data Engineer / Analyst

Brisbane QLD · 0400 000 000 · [email protected] · github.com/ellamarchetti · linkedin.com/in/ella-marchetti

Professional summary

Information technology and business graduate with a data specialisation, six months of commercial analytics experience, and a habit of checking the denominator before sharing the number. Strong SQL, working dbt and Airflow exposure, and enough business background to ask what a metric is going to be used for.

Technical skills

SQL: Advanced - window functions, CTEs, query tuning, Snowflake and PostgreSQL
Python: pandas, NumPy, matplotlib, requests, openpyxl
Data engineering: dbt (models, tests, docs), Apache Airflow (basic), Fivetran, Git
Warehousing: Snowflake, PostgreSQL, dimensional modelling (Kimball basics)
Visualisation: Power BI (DAX basics), Tableau, Excel to an advanced level
Statistics: Regression, hypothesis testing, time series decomposition (university level)

Education

Bachelor of Information Technology / Bachelor of Business Management (dual degree), Data Analytics major

Feb 2023 - Nov 2026

The University of Queensland

GPA 6.0 / 7
Relevant courses: Database Systems (7), Data Analytics (7), Statistical Modelling (6), Information Systems (6), Managerial Accounting (6)
Capstone: demand forecasting for a not-for-profit food relief service. Model reduced weekly over-ordering by an estimated 12% in a four-week trial

Experience

Data Analytics Intern

Nov 2025 - Feb 2026 (12 weeks)

Fernhill Insurance Group, Brisbane

Rebuilt the weekly claims report in Power BI, cutting preparation from 6 hours of manual Excel work to a 10-minute refresh
Wrote 22 dbt models with tests for a claims mart, including the first uniqueness and freshness tests the team had
Traced a persistent discrepancy between two claims reports to a duplicated broker record, which had been overstating one region's claim count by 8% for around a year
Presented findings twice to a non-technical operations forum of 15 people

Business Analytics Assistant (casual, 12 hrs/week)

Mar 2025 - present

UQ Student Services

Automated a monthly participation report from a manual spreadsheet process into a scheduled Python job with a documented data dictionary
Built the SQL views three staff members now use directly instead of requesting extracts

Assistant Manager (casual, then part-time)

Feb 2022 - Jan 2025

Riverbend Cafe, Brisbane

Managed rosters and stock ordering for a team of nine while studying full-time
Introduced a simple sales-by-hour tracking sheet that cut weekly food waste by around 15%

Projects

Queensland electricity demand explorer

Python, dbt, DuckDB, Streamlit, public AEMO data

Ingested five years of public half-hourly demand and price data, modelled it in dbt with tests, and published an interactive explorer
Documented three data quality issues in the raw feed, including a daylight-saving duplication that silently added 48 rows twice a year
github.com/ellamarchetti/qld-demand

Food relief demand forecasting (capstone)

Python, scikit-learn, statsmodels

Compared a seasonal naive baseline against regression and gradient boosting; the simplest model that beat baseline was chosen deliberately over the most accurate one
Delivered a one-page instruction sheet so volunteers could run it without me

Leadership and activities

Treasurer, UQ Data Science Society, 2025 - 2026. Managed a $9,000 annual budget and reported to a committee of eight
Volunteer data support, Brisbane community food relief service, 2024 - present

Certifications

dbt Fundamentals, 2025
Microsoft Power BI Data Analyst Associate (PL-300), April 2026

Referees

Available on request.

Why this CV works

The best bullet on this CV is a bug

Tracing an 8% overstatement to a duplicated broker record is precisely what a data team is buying. It shows the candidate noticed a discrepancy, chased it to a root cause, and understood the business impact - all three.

Time saved is the currency of analytics

6 hours to 10 minutes. Extracts replaced with self-serve views. Manual spreadsheet to scheduled job. Every analytics hiring manager is drowning in manual reporting; show them you remove it.

The public-data project is aimed at the employer

Using public AEMO electricity data when applying to an energy retailer is a deliberate signal. One project chosen for the industry you are applying to beats five generic ones.

Choosing the simpler model is a strength

Stating that the simplest model to beat baseline was chosen over the most accurate one shows judgement. Graduate CVs usually claim the fanciest technique; teams want the one that gets maintained.

Hospitality management is not a filler entry

Rosters, stock ordering, a team of nine, and a waste reduction driven by simple tracking. That last bullet is an analytics story told through a cafe. Reframe non-technical work by what it demonstrates.

SQL is listed first and claimed as advanced

For a data role, SQL is the skill that gets tested hardest and mentioned least confidently. If you are strong, say so and be ready to prove it live. If you are not yet, put the hours in before you apply.