Senior Machine Learning Engineer

Our client is looking to bring on a Senior Machine Learning Engineer to help build and scale a nextgeneration voice-centric AI platform used by millions. In this role, you'll own the full ML lifecycle-from research and data pipelines to training, deployment, optimisation, and evaluation-while pushing the boundaries of realtime conversational AI.

You'll focus on creating highly optimised, domain-specific conversational models, finetuning and compressing existing LLMs and speech models to deliver subsecond, natural interactions at scale. This is an opportunity for someone who has taken STT/TTS/LLM dialogue systems from prototype to production and thrives in fast iteration, deep optimisation, and real-world impact.

What You'll Do

• Build and optimise STT, TTS, and LLM pipelines that sustain < 800 ms latency across thousands of concurrent voice sessions.
• Apply LoRA, QLoRA, DPO, RLHF, quantisation, pruning, and distillation to create smaller, faster, cheaper conversational models.
• Finetune and evaluate lightweight architectures (e.g., DeBERTa, SetFit, sentence transformers) for intent classification and dialogue management.
• Optimise inference on GPUs/CPUs using dynamic batching, speculative decoding, and streaming gRPC/WebSockets.
• Develop data pipelines (Kafka Snowflake/S3) and build evaluation frameworks for ASR/WER, TTS MOS, and taskcompletion metrics.
• Implement and extend dialogue flows using LangGraph/LangChain and Model Context Protocol (MCP).
• Establish ML best practices: versioning, monitoring, CI/CD for models, safety guardrails, and production reliability.
• Work closely with product, infra, compliance, and customer teams to deliver safe, scalable conversational experiences.
• Stay on the cutting edge of speech + LLM research and rapidly prototype new methods (e.g., Whisper-v3, Bark, efficient finetuning techniques).
• Ensure HIPAAgrade security, PHI-safe handling, and always-on reliability.