Best suited to junior candidates with Python and SQL proficiency who are comfortable with ambiguous data and can communicate findings to non-technical stakeholders.
Derived from job-description analysis by Serendipath's career intelligence engine.
Original posting from Calibrax via Ashby
About the Role
Calibrax AI delivers end-to-end AI solutions, from strategy and model development to deployment, integration, and long-term optimisation. Our engineers build the systems that make AI actually work at scale: pipelines, APIs, integrations, and the infrastructure that keeps everything running. We work across a diverse range of clients and enterprises, in environments where reliability and performance are not optional.
We are looking for a Data Science Intern who can support our engineering and delivery teams in preparing, exploring, and modelling the data that feeds our internal systems and tooling. During this internship, you'll be working with real datasets, real constraints, and real deliverables, alongside consultants and engineers who expect you to bring rigour, not just curiosity.
Who You Are
- Currently pursuing or recently completed a degree in Data Science, Computer Science, Statistics, Engineering, or a related quantitative field
- Comfortable with ambiguity as client data can be messy, priorities shift, and you won't always get a clean brief
- Genuinely interested in how AI systems behave in production
- Able to explain your reasoning and your results to people who are not technical
- Self-directed enough to move with minimal hand-holding, but knows when to flag a blocker early
Must Haves
- Proficiency in Python for data work (pandas, numpy, and at least one ML library such as scikit-learn)
- Working knowledge of SQL
- Solid grounding in core ML concepts like regression, classification, clustering, and how to evaluate a model honestly
- Experience with data cleaning, exploratory data analysis, and visualisation
- Strong written English, since findings get communicated to the team and leadership, not just left in a notebook
What Success Looks Like
- Within the first few weeks, you're pulling, cleaning, and validating datasets independently, with minimal supervision
- You've contributed to at least one deliverable, a model, a dashboard, or a pipeline component
- You can walk a non-technical internal stakeholder through your analysis and its limitations without oversimplifying or overselling it
- You flag data quality or scope issues proactively, before they become someone else's problem
Source: Calibrax careers (Ashby)