Data Analyst Intern

Overview

Company Overview

Meshy is a Silicon Valley–based company and a global leader in 3D generative AI. We are redefining the 3D content creation workflow through cutting-edge AI technology. Our platform enables both professional creators and hobbyists to effortlessly generate high-quality 3D assets — transforming text or images into detailed 3D models within minutes. Meshy is fundamentally improving the efficiency and accessibility of 3D creation.

Our team includes top talent from MIT, Stanford, UC Berkeley, Nvidia, Microsoft, and other leading institutions and technology companies. With team members across North America, Asia, and Oceania, we foster a highly diverse and globally collaborative innovation culture. Meshy is backed by top-tier venture capital firms including Sequoia and GGV, with $100 million raised across Series A and

B. According to the A16Z Games 2024 report, Meshy is the most popular 3D AI tool. Based on Similar Web data, the platform receives over 4 million monthly visits. We have more than 7 million registered users and have generated over 70 million 3D models to date.

Founder and CEO Yuanming (Ethan) Hu holds a PhD from MIT in computer graphics and AI. He is the creator of the open-source GPU programming language Taichi (27K+ Git Hub stars, adopted by 300+ research institutions). His research has been nominated for the SIGGRAPH Best PhD Thesis Award and has received 2,700+ academic citations.

We are hiring a Data Analyst Intern to own two integrated pillars of our analytics work:

• Dashboard migration — moving existing visualizations from Databricks, Metabase, and Tableau onto Hex, our new AI-powered analytics platform.
• Analysis and proactive insights — running ad-hoc analyses, investigating anomalies you uncover along the way, and producing business insights that leads to key business decisions.

The pillars feed each other: as you migrate dashboards, you will notice anomalies worth investigating; as you investigate, you will sharpen the metric definitions and chart selections that go into the new dashboards.

This is a real, cross-functional project touching every business team  will get end-to-end ownership across the full scope above: scoping requirements with stakeholders, rebuilding dashboards in Hex with AI assistance, validating data parity, investigating what the data is telling you, and handing off polished work to stakeholders. The technical bar is intentionally accessible (the underlying data is already there), but the role demands attention to detail, patience, strong communication, and a genuine appetite for learning new AI tools.

This is not a ticket-execution role. We expect you to think critically throughout: which charts are obsolete? Which metrics could be defined better? Why is this segment behaving differently? What does Faye or the Growth team need to know that they haven't asked yet?

• Dashboard Migration & Data Accuracy
• Audit existing dashboards on Databricks, Metabase, and Tableau and define migration priority.
• Rebuild visualizations on Hex with exact data parity to the original.
• Design and run validation procedures (row counts, aggregate checks, sampled dimension comparisons).
• Improve on the new dashboards with new metrics and visualizations.
• Maintain a clear migration tracker (status, owner, blockers).
• Requirements Refinement & Metric Optimization

Before Migrating Each Dashboard, Proactively Engage Stakeholders

• Which charts and metrics are still actively used? Which can be deprecated?
• What additional dimensions or metrics would they want in the new version?
• Identify opportunities to enrich data and sharpen metric definitions so business teams can pinpoint issues more accurately. Push for standardized, documented metric definitions.
• AI & Python Tooling
• Leverage Hex's built-in AI to auto-draft chart skeletons and speed up migration.
• Use Python (pandas, plotly) for complex visualizations that AI cannot generate out-of-the-box.
• Explore modern interactive patterns on Hex (parameterized notebooks, AI-assisted exploration). Aim to make the new dashboards better than the old ones, not just equivalent.
• Validate every AI-generated output.…