Founding Engineer, Head of AI

Onlykit is a VC backed startup (led by Context VC & Hustle Fund) building a conversational voice AI for defense and public safety. We’re hiring a founding engineer to develop a conversational voice AI for use as a tactical copilot in various frontline use cases. As founding engineer at Onlykit, you will design and train LLMs and CSMs, and work with the engineering team to build an end to end speech pipeline. You will also collaborate on deploying our voice AI on AWS GovCloud and on self-hosted edge compute devices. In addition, you will be responsible for designing a test suite to benchmark the various speech models that power Onlykit’s voice AI. The role is remote anywhere in the United States, requires U.S. citizenship or permanent resident status (green card), and a background check.

What we need you to bring

• Real-time speech generation experience— experience implementing or maintaining low-latency TTS libraries.
• GPU model tuning skill—comfortable quantizing, pruning, and distilling large parameter models.
• Deep Python & PyTorch know-how—able to dive into model internals, and help optimize performance. Bonus if you are comfortable in C++.
• Audio-pipeline savvy—know how to slice live microphone or radio input, remove noise, stitch frames without pops, and keep word boundaries intact for clean playback
• Secure cloud deployment—skilled with Docker plus an IaC tool (Terraform, CDK, or similar) to launch GPU services in a private VPC, with encryption, patching, and audit logs
• Performance testing habit—write scripts that capture latency, throughput, and mean-opinion-score
• Clear documentation—you write straightforward notes and diagrams

If you’ve rebuilt a high-quality speech system before and can show benchmarks, let’s talk!