About the role
We are looking for a hands-on Lead AI Engineer to build and optimise Speech AI and Clinical AI solutions. The ideal candidate should have deep expertise in fine-tuning LLMs and ASR models, model evaluation, distributed training, and deploying production-grade AI systems.
Responsibilities:
- Fine-tune Speech AI (ASR) and LLM models for clinical applications.
- Build training datasets and evaluation pipelines.
- Optimise models for latency, scalability, and production deployment.
- Lead technical decisions, mentor engineers, and collaborate with Product, Clinical, and MLOps teams.
- Design robust AI systems using fine-tuning, RAG, and distributed training techniques.
Required Tech Stack:
- Programming: Python, SQL, Bash.
- ML Frameworks: PyTorch, Hugging Face (Transformers, PEFT, TRL, Accelerate).
- LLMs: Llama, Qwen, Mistral, Gemma.
- Fine-Tuning: LoRA, QLoRA, SFT, RLHF, DPO, ORPO, PPO.
- Speech AI: Whisper, wav2vec2 HuBERT, Conformer, NeMo, ESPnet, SpeechBrain.
- RAG and NLP: RAG, Vector Databases, Prompt Engineering, Clinical NLP.
- Distributed Training: DeepSpeed, PyTorch FSDP, Multi-GPU.
- Inference: ONNX, TensorRT, vLLM, Triton Inference Server.
- Cloud and MLOps: AWS/GCP/Azure, Docker, Kubernetes, MLflow/W& B, Git, CI/CD.