(Incubator) Stealth Startup: Deep Learning Engineer
Are you interested in delivering a deep learning product to an early list of excited customers?
Do you want to help develop and deploy state-of-the-art deep learning models?
If so, then we’d love to talk to you. Our team has used the latest breakthroughs in deep learning to create a text-to-speech product that sounds truly humanlike. We’ve invented new neural architectures to recreate human speech and we are looking for you to help us systematize this state-of-the-art product to fulfill the needs of an early list of customers.
You’ll be responsible for productionizing deep learning models, from data ingestion to model optimization through model deployment. You’ll implement and enforce great engineering practices while also having the opportunity to experiment and help write the playbook on how to deploy and scale entirely new technology. You’ll get to work with PyTorch and prototype ideas quickly.
Given how new this technology is and how quickly startups and markets can evolve, you will benefit from being curious, flexible, and motivated to figure out hard stuff.
What you’ll do:
- Write low-level optimized CUDA kernels to run our models efficiently on top of a distributed infrastructure
- Build a distributed backend with many GPUs allowing us to reasonably scale our product to hundreds, then thousands, of concurrent users
- Build out high quality datasets to help scale and refine our text-to-speech services
- Refine our model training process, ensuring model reliability
- Manage a growing team of backend engineers and deep learning engineers
- Contribute back to the community through papers and open source code
Experience you have:
- Python and C / C++ experience
- Familiarity with PyTorch or Tensorflow
- Experience deploying models via GCP or AWS
Finally, we’d love to show you a demo if you swing by the office.