Data Science Intern
Partly is connecting the world's parts, and we're doing that by building the AI infrastructure layer for the global repair industry, starting with the $2tn automotive market. Our frontier model, Interpreter, is the world's first AI purpose-built to understand vehicle damage and the parts needed to fix it. Thousands of businesses across the global repair supply chain already rely on it.
Founded by ex-Rocket Lab engineers, we've tripled in size in the last 18 months and have recently raised a $50m Series B led by DST Global (Anthropic, Airbnb, Meta, TikTok, Spotify) and including Blackbird Ventures (Canva, CultureAmp etc.), WNDR, Activant Capital, Icehouse Ventures, Square Peg, Airtree, and Ecliptic Venture Capital. We're headquartered in Austin, with offices in New Zealand and London.
As a Data Science Intern, you'll work alongside our Applied ML and DataQA teams to help make sense of the technical problems behind Partly's core vehicle and parts workflows. You'll be paired with experienced data scientists and engineers who will mentor you as you dig into messy, real-world data and turn it into clear, useful analysis.
The work sits between data science, domain analysis, and product. You'll help investigate why things like parts validation, parts ordering, automatching, and variant handling become difficult or unreliable, and you'll help quantify how often those problems happen and how much they matter. This is not dashboards for the sake of dashboards, and it is not pure model training. It's hands-on, investigative work on real problems with real impact.
This is an internship for someone early in their journey who learns by doing. Expect to work directly with SQL and Python, ask good questions, build lightweight tools or analyses, and see your findings help the team decide what to fix next.
What You Will Do
Dig into real parts data. Help build a data-backed understanding of the problems that make parts validation, parts ordering, and automatching harder, with guidance from your mentor.
Help quantify problem types. Assist in breaking broad issues (for example, "variants are messy") into specific, measurable categories and counting how often each one shows up.
Measure where quality is lost. Use SQL, Python, sampling, and internal tools to help estimate accuracy, failure rates, and impact across makes, providers, and part groups.
Turn analysis into something useful. Help produce clear findings that Applied ML, Product, and DataQA can actually act on.
Build lightweight tooling. Pitch in on scripts, small dashboards, and review workflows that help others see and understand parts problems more clearly.
Learn how the team works. Partner with Applied ML, DataQA, and Product teams, and learn how a high-velocity, low-bureaucracy team operates.
Get the basics right. Produce reproducible analyses, write clear notes, and ask thoughtful questions that make your work easy to build on.
Your Skills
Solid analytical fundamentals. You're studying or have recently studied data science, statistics, computer science, mathematics, engineering, or a related field, and you can reason carefully about data.
Working SQL and Python. Through coursework, projects, or prior experience, you can query data and write code to analyze it. You don't need to be an expert, but you should be comfortable getting your hands dirty.
Good problem decomposition. You enjoy taking a vague or messy question and breaking it into smaller, answerable pieces.
Curiosity about how things really work. You like understanding the "why" behind a problem, not just producing a chart, and you're comfortable with data that isn't clean or fully labeled.
Clear communicator. You can explain what you found and why it matters, ask for help when you need it, and take feedback well.
Bias for learning. You want to be stretched, you take ownership of your growth, and you're excited to work on hard, real-world problems rather than tidy textbook ones.
(Bonus) Any exposure to classification problems, data quality or QA work, entity matching, catalog data, automotive data, or working alongside ML or human-in-the-loop systems.
Please note: if you don't have all the skills or experience listed above but believe you could be outstanding in this role, please still consider applying. Many people count themselves out. We'd love the chance to learn more about you and why you're exceptional.
Relocation
If you are relocating from overseas or domestically to Partly HQ, we offer a generous relocation allowance to support your move