VoiceOps uses AI to improve call center rep performance with world-class coaching.
Our average customer makes tens of thousands of calls per week. In a world without VoiceOps, they have literally no idea what their sales reps are doing on the phone. It's a total (and scary) black box.
By applying ML and a great UI to this problem, call center leadership has all the data they need about customer conversations at their fingertips, and can coach their reps more effectively and efficiently.
The technical problem is interesting, and gets more interesting as we grow. Our core challenge is how to take billions of audio recordings (and messy, unstructured human conversations) and make sense out of that data in a way that is: a) accurate b) cost efficient, and c) highly scalable. The corresponding product problem is how to take well-structured data and make it actionable for the end-user.
Call center recordings are one of the richest/largest untapped datasets in the world (literally, billions of calls stored in AWS buckets that no one is touching right now). We're going to be the best in the world at structuring that data and putting it to use to make businesses work better.
Data Platform Engineer San Francisco, CA, United States, Silicon Valley, CA, United States, or Remote
Senior Back-End Engineer San Francisco, CA, United States, Silicon Valley, CA, United States, or Remote
Why join us?
Impact — Because the team is small, you’ll help set the template for engineering quality and process. You'll also be encouraged to contribute to product decisions, and help shape the direction of the company.
Stellar Team and Culture We have very high standards for our engineering team, so you'll get to work with some of the smartest people you've ever worked with.
Growth — We've just raised our Series A and are looking for more smart people to continue growing quickly. You’ll be one of the first few employees and have opportunities to be a leader on a growing team.
Engineering at VoiceOps
We currently have 9 engineers, including our Head of Engineering, divided across front-end, back-end, infrastructure, and machine learning. The team currently works in 2-week sprints, each organized the Friday before start. These sprints are planned collaboratively with large projects scoped, triaged, and estimated by the team themselves - large features are assigned an engineering lead on a rotating schedule who is responsible for speccing out the project from an engineering standpoint.
The team relies on Agile methodology and Jira ticket tracking systems to keep themselves organized. As part of an internal commitment to early engineering excellence, we hold bi-weekly retrospectives to iteratively improve our processes and continually optimize for further efficiency and effectiveness.
We optimize for tight execution on a limited set of priorities, which creates a collaborative environment (we're often all working together on the next big user-facing project together) and mutual accountability.
Human conversations are messy. Previous software attempts at structuring conversation data leave a lot to be desired.
We are taking the same approach to conversation analysis that Uber/Cruise/Waymo are taking for self-driving cars. Building troves of training data, and solve lots of edge cases piece by piece towards the goal of having an incredibly reliable system.
Some upcoming challenges we anticipate are:
Optimizing our database usage at scale - We intend to hire 2 account executives per month for the foreseeable future, and with a larger sales team constantly out winning new business for VoiceOps we need to make sure that our data processing and quality doesn’t suffer as we scale. Moving from processing millions of calls to billions of calls will certainly be a challenge, but one we’re excited to face.
Splitting our monolithic web app into microservices - Our current codebase is a monolith. As we scale, we’d like to migrate to microservices, which will serve us better in the long run.
Gamifying audio training for new transcribers - The current onboarding process for new transcribers is clunky, and as we scale retaining those transcribers will be key. By gamifying the training piece, we hope to increase retention / minimize dropoff of these new transcribers.
Improving ASR outputs - We have a lot of data on which words and phrases are frequently corrected between our ASR text and our final clean transcription. This data constitutes some prime training data for a simple model that auto-corrects common ASR errors. Get ready to tokenize a majillion n-grams, and save tons of VoiceOps time and dollars by making transcription better.
Implementation of a robust, configurable, thread-safe state machine to move calls through various analysis steps, offering visibility if any of the steps are slow or failing. Before this project, calls could go through the same step multiple times or be processed out of order due to stale object references or multiple threads executing at once.
An integration system which imports call recordings from dozens of customer systems, supporting the analysis of thousands of calls per day. Challenges include failure visibility, networking/VPN constraints within customer data centers, and careful handling of timezones / date cutoffs with international customers.
Break out bounded services from an existing monolithic application to allow for independent deployability and scalability of key components. Integral to success are : finding the appropriate seams on which to split various modules, containerizing discrete services, and choosing appropriate data access patterns.
Working at VoiceOps
Every problem is our problem
We do not look to external sources for why we didn't hit a deadline, meet an objective, etc. We hold ourselves responsible.
Intellectually honest and curious
We challenge each other's ideas daily - on product strategy and beyond. At lunch we are more likely to talk philosophy than TV (though we do some of that too)
Emotionally open and vulnerable
We talk about the ups and the downs of our lives outside of work and strive to have high authenticity interactions with each other.
We work hard to be an enduring company that paves the way for all sorts of applications which require structuring verbal conversational data at scale.
High quality bar, and always rising
We redo our work again and again until it's great. (can be frustrating in the moment but is ultimately extremely rewarding when we build things that are greater than we imagined they would be)
Work from Home
Interested in this company?
Skip straight to final-round interviews by applying through Triplebyte.