Machine Learning Engineer – Real‑time Voice AI

About the role

We are seeking a Machine Learning Engineer to design, deploy, and operate real‑time speech AI services for a high‑scale production platform. The role is fully remote and focuses on delivering low‑latency, reliable inference across automatic speech recognition, translation, and text‑to‑speech pipelines.

Key responsibilities

• Deploy and optimise ML models for ASR, translation, and TTS in production.
• Build low‑latency inference services using FastAPI or similar frameworks.
• Maximise GPU utilisation, throughput and p95/p99 latency through batching, caching and inference‑optimisation techniques.
• Integrate and tune technologies such as TensorRT, Triton, TorchServe, vLLM and CUDA.
• Profile, debug and resolve performance bottlenecks across real‑time pipelines.
• Maintain and monitor production ML systems serving real users at scale.

Required profile

• 4+ years of experience in ML engineering or ML systems.
• Proven track record deploying ML models to production.
• Hands‑on experience with speech‑related NLP, ASR, translation or TTS.
• Strong focus on performance, reliability and scalability.

Required skills

• Python
• FastAPI (or similar API framework)
• TensorRT
• Triton Inference Server
• TorchServe
• vLLM
• CUDA
• GPU inference optimisation
• ASR, translation, text‑to‑speech technologies
• Familiarity with Whisper, NLLB or other multilingual models (nice to have)

What we offer

• Opportunity to work on a high‑impact AI modernization project.
• Modern tech stack with challenging engineering problems.
• Ability to shape architecture and make key technical decisions.