Staff Voice AI Engineer - Applied AI

About This Role

Lead the design and deployment of large-scale, real-time Voice AI systems to power next-generation product experiences at Uber. This role involves full-stack technical leadership across speech modeling, LLM-powered conversational intelligence, and low-latency backend infrastructure.

Responsibilities

• Design and build end-to-end Voice AI solutions from understanding customer needs to deploying LLM-powered, real-time voice interfaces in production
• Benchmark and evaluate voice AI systems, including speech recognition, speech synthesis, and spoken language understanding, by designing evaluations, analyzing results, and identifying systematic weaknesses
• Improve voice model performance through system prompt tuning, fine-tuning voice- and speech-specific models, and optimizing architectures for low-latency, real-time voice interactions
• Analyze voice request logs, prompt traces, and audio inputs to diagnose failure modes, improve transcription accuracy, conversational quality, and overall user experience
• Build and maintain internal tools and platforms to automate Voice AI workflows, such as large-scale transcription pipelines, real-time audio processing services, and evaluation harnesses for voice quality
• Own Voice AI systems in production end-to-end, including rollout strategies, monitoring, alerting, quality regression detection, and on-call readiness
• Collaborate closely with product, design, and research teams to translate user needs into Voice AI capabilities with measurable business and customer impact

Requirements

• 10+ years of experience in software engineering, data science, or machine learning
• Track record of shipping production AI systems
• Deep understanding of large language models (fine-tuning, prompt engineering, embeddings, RAG)
• Strong backend and distributed systems expertise, with experience designing and operating highly available, scalable services in production
• Deep experience with ML infrastructure (model training pipelines, online serving systems, feature stores, experiment platforms, evaluation frameworks)
• Hands-on experience with distributed data processing systems (e.g., Spark, Flink, Ray) and workflow orchestration (e.g., Airflow or equivalent)
• Ability to analyze data, run experiments, and derive insights for model and product improvement
• Excellent communication and collaboration skills

Qualifications

• 10+ years of experience in software engineering, data science, or machine learning, including a track record of shipping production AI systems

Nice to Have

• Experience building evaluation frameworks for Voice AI (metrics, human/LLM-assisted evaluations for speech recognition accuracy, latency, robustness, and naturalness)
• Demonstrated expertise in machine learning fundamentals applied to voice (model evaluation, training, and fine-tuning of ASR, TTS, or speech-language models)
• Proven experience deploying Voice AI systems to production, with an emphasis on low-latency, high-reliability, real-time environments
• Experience writing developer documentation, creating voice-specific SDKs, or enabling internal teams to build on shared Voice AI platforms
• Hands-on work with large-scale audio datasets, including data curation, labeling strategies, and optimization of voice processing pipelines at scale

Skills

* Required skills

Benefits