Machine Learning Engineer (TTS, STT)

Our client is on a mission to democratise AI‑powered voice agents — enabling businesses worldwide to automate customer and operational calls with intelligent, natural-sounding dialogue. With over 30 million automated calls handled, 4 million hours saved in contact centres, and 99.9% uptime, their no-code platform is powering the future of voice AI.

Backed by the investors behind Facebook and GroupOn and trusted by 1,000+ companies, they enable enterprises to deploy multilingual, human-like voice agents at scale.

The ML Engineer will join the founding team, shaping real-time, low-latency speech systems from the ground up. This is a hands-on, architecture-level role - building end-to-end voice pipelines that combine ASR, LLMs, and TTS, and scaling them across mission-critical environments.

• Design and deploy real-time ASR → LLM → TTS pipelines optimised for speed and naturalness
• Fine-tune and integrate best-in-class open-source and commercial models across ASR, TTS, and LLMs
• Optimise inference performance using techniques like quantisation, distillation, and RL-based tuning
• Build high-throughput, low-latency APIs using Python, FastAPI, gRPC, and WebSocket streaming
• Own deployment workflows across cloud and on-prem, collaborating on Docker, Kubernetes, and CI/CD
• Stay ahead of the curve — run experiments, track new research, and help keep the team at the cutting edge

• Deep experience in Python.
• Proven hands-on work with real-time ASR, TTS, voice chat, or streaming audio systems
• Familiarity with LLM tooling - prompt design, fine-tuning, RAG, and frameworks like LangChain, LlamaIndex, or OpenPipe

You’ll push the boundaries of what voice AI can do in production - balancing model quality, latency, and scalability - while working across product, infra, and research to ship new capabilities into the hands of real users.

If you want to build the voice layer of enterprise AI - and work at the frontier of real-time language, speech, and systems - we’d love to hear from you.