amauboussin

Surge AI | Full-Stack Engineers | Remote | Full-Time or Contract | https://surgehq.ai

AI is going to change the world, but the advances of the past few years have been slow to translate into real-world impact. Our mission at Surge is to change that.

We’ve been developing two products: A platform for extremely high-quality data labeling. Think Mechanical Turk 2.0, but with an elite workforce: in order to build models of misinformation and hate speech, for example, we need to leverage skilled workers who understand complex problems. Tools for easily turning data into AI. A machine learning model is defined by its data, so anybody -- from your PM to your marketer to your engineers -- should be able to create ML models just by labeling data points.

We’re a team of ex-Google and Twitter engineers who’ve quickly bootstrapped into very healthy revenue. Reach out to team@surgehq.ai if you’re interested in chatting! Our tech stack is Ruby on Rails, React, and Python, and we’re looking for backend, frontend, and ML engineers, but love talking to talented folks of any types. (We also love personal projects and writings if you have any to share!)

Surge AI | Remote or New York City | Full-time

We believe the biggest blocker to most ML use cases is high quality training data so our first product is a platform for obsessively high-quality training data (see our blog for more details https://medium.com/swlh/we-need-new-infrastructure-for-the-n...). The end goal is to build a platform that makes it easy to get high-quality data at scale -- think of labeling 100K tweets for misinformation in order to build a fake news algorithm, or annotating 500K images to measure AI fairness -- and then instantly turning that data into AI. We are looking for product engineers (entry level, team lead, or anywhere in between) to join the team and help us develop our platform. It is built on Ruby on Rails, React, and Python (for the machine learning components).

A few example projects you could take on: * We work on many complex, subjective tasks like hateful speech. Is “stfu trump” toxic content that should be downranked and hidden, or is it a valuable voice that speaks truth to power? In order to maintain quality, we need to think about systems that train workers (like interactive quizzes or videos) and monitor the quality of their work. * We often label hundreds of thousands of tasks each day. At this scale, efficiency becomes important. How do we automatically optimize layouts and give agents better tools to work efficiently? Can we integrate ML to help automate some of the work? * We want to make building AI as easy as building a spreadsheet. How do we make this a fun, understandable experience? Can we give you automatic insight into your data and AI models?

Email me if you are interested! andrew@surgehq.ai