Google’s internships are an important part of our culture of building for everyone. Internships are designed to be more than just a summer job; it’s an opportunity to tackle real-world challenges and make an impact. Interns work alongside full-time Googlers, contributing to the helpful products and services that people use every day. While they gain hands-on experience and grow their skills under the guidance of dedicated mentors, we benefit from their curiosity and new approaches to problem-solving. To learn more, visit our Google Careers site, set up job alerts, and apply when applications open in the fall.
Q. Meet our intern Daksh Tulsyan. Tell us about yourself!
I’m a fourth-year undergraduate student at NIT Trichy studying Computer Science and Engineering. This summer, I had the fantastic opportunity to join the Data Commons team as a Software Engineering Intern. I’m excited to share a glimpse into my internship and the open-data feature I had the privilege to build. It’s been an incredible journey of exploring the knowledge graph, collaborating across teams, and bridging the gap between open data and machine learning.
Q. What opportunity were you solving for?
Data Commons aggregates one of the world’s largest knowledge graphs of public statistical data. A major issue we wanted to address was a functional gap: the absence of standardized, machine-readable metadata to achieve global data discoverability.
Without a standardized footprint, our massive data inventory was largely invisible to global discovery engines like Google Dataset Search. Furthermore, AI agents and automated ML pipelines couldn’t easily discover or ingest our datasets, as they lacked the machine-readable structures necessary for automated consumption.
My project focused on closing this gap with Croissant, a high-level format for machine learning datasets built on schema.org and serialized as JSON-LD. By providing a standardized vocabulary to describe our datasets, we could make our data instantly AI-ready and broadly discoverable.
Q. What is the solution you helped implement?
We implemented a dynamic pipeline to generate and expose Croissant metadata in JSON-LD format for datasets across Data Commons. Here is a look at how this integration transforms the Data Commons knowledge graph browser flow:
The dynamic Croissant generation pipeline intercepts graph browser requests, compiles metadata directly from dataset, source, and provenance properties via the API server, and injects compliant JSON-LD payloads.
When a user navigates to a dataset node on our graph browser (e.g., via /browser/<dcid>), our backend Flask controller intercepts the request, verifies the node type, and queries the API to fetch the required dataset and source properties. Following this, it executes a graph traversal (<-isPartOf) to discover all of the underlying provenance nodes, as modelled in Data Commons, and the related properties linked to the dataset.
The backend then enforces Croissant compliance, cross-checking for required fields like name, url, license, datePublished, and creator, while ensuring description texts fall within the MLCommons 50-5000 character limit. If the requirements are met, we bundle the metadata into a Croissant JSON-LD payload and safely inject it into a <script type="application/ld+json"> block in the HTML <head>. To help users navigating Data Commons, I implemented an inline “Explore dataset” button on the UI that renders when this rich JSON-LD data is successfully mapped behind the scenes.
With this structure in place, the datasets can now be crawled by search bots and indexed directly on Google Dataset Search. Users can navigate to the datacommons.org portal to learn more about the dataset, explore the data for coverage on topics of interest, time and places.
Q. What are some of the key learnings you are taking away from this internship, whether project or experience-related?
Having the opportunity to work at the intersection of open data and machine learning has been an incredible journey of technical learning and cross-functional collaboration. This internship taught me how to shift from building isolated academic projects to deploying features at scale. Understanding how search engines parse metadata in the real world completely changed my perspective on open data. More importantly, I learned that shipping impactful features is deeply collaborative. Getting direct feedback on schema mapping from the creator of the Croissant standard, Omar Benjoillian, was a great experience.
Additionally, getting hands-on with tools like Antigravity and the power of AI-assisted development to accelerate engineering workflows. Moving forward in my university curriculum, this experience has inspired me to focus on macro-level system design and data standardization—looking beyond just core programming to understand how complex, interoperable architectures are built. Ultimately, knowing my work opens up Data Commons to global search engines and AI tools has been incredibly rewarding!