Labelbox provides a collaborative, extensible platform, to create and manage training data at scale. We help our customers make machine learning their competitive advantage.
The biggest problem today for companies creating a machine learning model is not coding or training the model it's getting a high quality dataset. That's where companies are getting stuck and spending 90% of their time.
We have over 50 paying customers (multiple fortune 500 companies) and would be profitable by the end of the year if we stopped hiring
Labelbox raised $3.9 million dollar seed at 5 months old and a $10 million dollars at 12 months old. Our investors include Kleiner Perkin, Gradient Ventures (Google's AI Fund), and First Round Capital.
The engineering team is currently 6 people and we operate on 1 week sprints. We have monthly OKRs and a 2 year roadmap. Each engineer is responsible for 1 OKR item and works with others in the team to design and develop that feature. Iteration and development is very fast due to team and company size.
Labelbox is providing the training data pipeline for companies all around the world. Each company has different types of data, different types of pipeline needs, varying data scale challenges and different number of collaborators. Building a reliable and flexible product for these organizations is very challenging and we need great software architects.
Given an image that takes a human 45min to label, what software tools can we give this labeler to speed up their job? Can we apply edge detection to outline existing objects? Can we train a model and pre-label the data?
How can we better develop data pipelines to help our users manage import / export of vast datasets? Providing webhooks to support multiple step labeling jobs? Developing bulk APIs to perform operations on huge datasets? And creating a python SDK to support our data science users
With large training datasets teams need strong collaboration features to automate data quality. Such as real time graphs on label agreement, commenting and requeueing, and slicing and dicing of complex data pools.
Make Customers Win
We find great pleasure in creating solutions for our customers. By working with us, our customers are at the forefront of industrial machine learning.
Seek to understand
Machine learning is growing faster than any other industry. We must stay humble and embrace that we have so much more to learn. We seek to understand first so that we build the most impactful solutions.
Craftsmanship is the fine balance between perfection and a solution to the given constraints. To be a craftsman is to be consumed in a problem and to create your best expressions of work as a result. You can't be a craftsman in every thing, you can only be a craftsman in a few things, but you do those things remarkably.
The team is going to Yosemite in July. I'm going to Peru in June. We have unlimited vacation and employees actually use it.
Lunch is provided and on Wednesday's we go out to eat at a restaurant. The mission where we work is an excellent food district.
We pass the
Burrito Ring to different team members who are in charge of scheduling something fun. We have a beach day coming up and have already done mini golf, bowling, and other fun team activities.
Interested in this company?
Skip straight to final-round interviews by applying through Triplebyte.