Labelboxer Spotlight: Nikki Gantos, Engineer

Meet Labelbox Senior Software Engineer Nikki Gantos in our first employee spotlight. Nikki has been at Labelbox for two years, and works on canvas tooling and API endpoints. Read on to learn more about her journey at Labelbox.

How did you become an engineer?

After a youth embracing the arts, I discovered engineering inadvertently at UC Santa Barbara while co-creating mutant bicycle contraptions. Shortly after, I spent a few years at Georgia Tech for a degree in Mechanical Engineering. During that time, I worked with robotics applications at the JPL, Oregon State’s RDML, and took a computer vision course. I began to explore software engineering too. After graduating, I joined DroneDeploy where I worked with image processing as well as full-stack development. While at DroneDeploy, I got to co-author a custom ThreeJSA view for their tiled drone map datasets optimized to render both point cloud and vector tiles. My focus since has been building the canvas tooling and API endpoints for the Labelbox platform.

Nikki Gantos at the Labelbox office in San Francisco.

Why did you decide to join Labelbox?

The people, particularly the founding members of the team, are what first brought my attention to Labelbox. What hooked me at Labelbox was the opportunity to build a product using a cutting edge tech stack and to be an early member of a growing team of engineers who share that vision. It has been exciting to help cultivate a team of incredibly bright ICs and managers.

Labelboxers on a team hike in 2019.

What makes engineering problems at Labelbox unique?

The incredible variety of the content available for annotation makes the engineering problems at Labelbox a nuanced and highly dynamic challenge. While many programs are tasked with  parsing a relatively constant input, the Labelbox platform has continued to expand horizontally to suit the needs of a variety of applications, from text to images to videos to tiled data. Inherent in this breadth is the need for a robust abstraction that the annotation platform can leverage to handle numerous and ultimately customizable canvas scenarios. As the company continues to expand this application, it becomes a fun abstraction problem to optimize both at the API level and the client. Doing so reliably at scale takes said abstraction to yet another level of complexity. In that way, I feel that the engineering problems at Labelbox are truly at the cutting edge of a novel industry application. We are a generation of pioneers in what may be the dawn of the AI age.

What have you learned during your time at Labelbox?

As a generalist at Labelbox, I’ve been able to work all over the stack. And in doing so, I’ve been able to develop a fluency in React, graphQL and even mySQL. Since much of the stack is in Typescript, it makes traversing between the API and client incredibly seamless. For a full-stack engineer, the mono-repo structure of the API and client really does take the barrier off of what is traditionally “frontend” or “backend” engineering. I think I am most delighted with my recent SQL prowess that allowed me to run a major database migration for our annotation features last quarter, something I would not have trusted myself with a year ago.

What gets you excited about coming to work every day?

The scale and diversity of the Labelbox engineering problems are what makes the role challenging, but at the same time, invaluably interesting. In the two years that I have been with the company, there has not really been a time where things became monotonous, and the novelty of opportunity here is what makes me most excited to work each day. There’s always something engaging to work through.

Which Labelbox value resonates with you the most?

Seek to understand. It really speaks to the nature of software development as an iterative process. And it is well attuned to the engineering solutions we are constantly evolving  at Labelbox to handle the complexity inherent in the variety of content available for labeling. But there is a sense of humility that resonates most with me in seeking to understand — that it’s okay to experiment, to question, and to be proven wildly incorrect — so long as the learnings inform a better outcome, enhance any initial assumptions, and we continue to grow. And there’s something fundamentally comforting in that.

Want to join the Labelbox team? Explore our open positions and apply today. Stay tuned in the coming months for more spotlights on Labelbox employees!



Labelbox is a collaborative training data platform empowering teams to rapidly build artificial intelligence applications.