About the job
Wingman is a product that helps salespeople sell better by using machine-learning to understand sales conversations.
Wingman was a part of YCombinator, Summer 2019.
It records sales calls and emails, uses machine-learning to identify trends and provide insights.
Sales managers use these insights to:
Wingman's customers are mostly sales teams of tech companies based in the U.S. Thousands of sales people are using Wingman today to get better at selling.
Consuming audio is inherently time consuming. As a side effect, people rarely use the insights hidden within audio data to improve their processes. Wingman uses cutting-edge machine-learning technologies to convert unstructured audio into actionable insights in real-time! This involves many challenging problems that require us to be at the forefront of infrastructure architectures and machine learning advancements to "understand" and influence customer conversations. Come talk to us to see what we've built and get a peek into the possibilities beyond.
About the role
This opportunity is for data scientists/engineers who have experience with the current state of the art in NLP techniques [e.g. Bert]. The process typically starts by working closely with the product team to convert a customer pain point into a problem statement [e.g. semantic search over transcripts, build chapters/index for each sales call, call summarization]. You will then be responsible for the end-to-end process of prototyping, evaluation and productionizing the solution. Some of this infrastructure is responsible for processing the audio data in real-time and in other cases post-call. Designing a system to process data in real-time is complex when your responses are invalid if they are delayed by more than a few seconds. You will be expected to also think like an engineer rather than purely as a data scientist.
What's in it for you