Research Conversations with Shijie (Brandon) Bian

Shijie (Brandon) Bian was first introduced to Autodesk during a stint working on a NASA-sponsored project. When the opportunity arose to intern with the company, Brandon knew he could expand his expertise in machine learning to tackle real-world problems at Autodesk. So, he spent last summer working on projects to help achieve his goal of providing users with automated applications that make an impact.

How did you learn about the internship opportunity with Autodesk Research? 

During my undergraduate studies at UCLA, I worked as a research assistant at the NASA-sponsored Autonomy Research Center for STEAHM (ARCS) on a project where we collaborated with Autodesk.  Despite being new to design automation, I was immediately intrigued by this project because it is so closely tied to users and customers. I worked closely with Autodesk Research engineers, and we applied an existing graph neural network model to the Autodesk Fusion 360 Assembly dataset. I published our work as the first author at the ASME IDETC Design Automation Conference and was thrilled to be able to continue my work with the company as a Research Engineer Intern.

What did you work on during your time as an intern at Autodesk? 

During my internship, I continued some of the work I did with Autodesk at ARCS. This primarily meant expanding the graph neural network model to fully capture geometrical information of CAD bodies before concatenating all features together and passing them through the global graph neural network for topological information learning. This can help designers by automatically providing material selection suggestions based on the knowledge learned from the hundreds of thousands of past designs in Autodesk’s design repository while greatly reducing the time and expertise.

What was your biggest takeaway from your internship at Autodesk Research? 

Quite simply, I learned that no person or project is an island. Our work is a combination of ideas from multiple realms, ranging from Computer Science (Machine Learning algorithm development), Mechanical Engineering (CAD design data and repository) to User Research (understanding what users need). I worked with a lot of researchers at Autodesk from multiple disciplines and could never have achieved this project without their help!

What was your favorite part of your internship? 

Everyone is humble and helpful at Autodesk. I can freely ask for help from senior engineers and principal researchers without any concern, and everyone is willing to help and contribute to our project. I also had the opportunity to meet several colleagues and supervisors during my in-person office visit and the IDETC conference – we had a lot of fun discussions on subjects beyond work!

Interns, like Brandon, have the opportunity to work on projects at Autodesk, publish their work, and can even present their findings at conferences. As we start to say goodbye to our summer interns this year, look to this space for more information on their research, work, and projects!