Description

You will fine-tune, deploy, and orchestrate LLMs for real-time speech-to-speech and multimodal voice agents.

This role is on-site.

Responsibilities

• Fine-tune and deploy LLMs for low-latency, high-accuracy voice interactions.
• Build and optimize inference pipelines for production-grade voice agents.
• Convert unstructured, noisy audio and text into clean training and evaluation datasets.
• Orchestrate LLM components, including fallback mechanisms and observability.
• Own the end-to-end model lifecycle: research, training, deployment, testing, and monitoring.

Required Skills

• 5-10 years of experience in AI/ML engineering or research.
• Strong proficiency in Python for model development and ML pipeline implementation.
• Hands-on experience with Speech Recognition (STT) systems for real-world audio.
• Proven track record of LLM fine-tuning and deployment (instruction tuning or pretraining).
• Experience with LLM integration and orchestration for production workflows.
• Ability to ship ML models with robust testing, iteration, and monitoring for reliability.

Preferred Skills

• Experience with multimodal interfaces and real-time conversation systems.
• Background in data processing for noisy, real-world audio/text sources.

Key Skills

Education

Any Graduate