Member of Technical Staff - Voice Model

Details

SpaceXAI's mission is to create AI systems that can accurately understand the universe and aid humanity in its pursuit of knowledge. The team is small, highly motivated, and focused on engineering excellence. Employees are expected to be hands-on, contribute directly to the company's mission, communicate clearly, and demonstrate strong work ethic and prioritization skills.

About the Role

Join the Grok Voice Model team to help build voice AI with smooth, natural, low-latency spoken interactions that are expressive, multilingual, and reliable across devices and real-time scenarios. The team owns the full training pipeline, including large-scale data curation, premium audio processing, speech-language model pre-training, and post-training focused on quality, speed, and stability.

Responsibilities

• Design and execute large-scale speech data curation and processing pipelines, including diverse real-world audio collection, synthetic data generation, and automated annotation workflows for model training and evaluation.
• Work on pre-training and post-training of speech-language models using supervised fine-tuning, reinforcement learning, and other techniques.
• Improve response accuracy, factual grounding, natural and idiomatic spoken style, conversational tone, and multilingual fluency.
• Build and iterate evaluation frameworks covering objective metrics such as accuracy, quality, latency, and expressiveness; human preference studies; content factuality assessments; real-time interaction quality; and experimentation infrastructure.
• Work with product teams to integrate voice models into applications and real-time environments.
• Define spoken interaction specifications and manage the lifecycle from prototype through global-scale deployment for stable, low-latency voice experiences.

Requirements

• Expert-level Python skills and deep proficiency writing clean, efficient code for AI and machine learning systems.
• Hands-on experience processing large-scale datasets using tools such as Spark and Ray for cleaning, augmentation, and feature extraction.
• Proficiency in pre-training and post-training speech-language models using JAX or PyTorch, including supervised fine-tuning, reinforcement learning, and optimization for accuracy, factuality, natural spoken style, detail, and multilingual fluency.
• Ability to establish and run rigorous evaluation pipelines involving objective metrics, human preference studies, content factuality checks, and iterative A/B testing.
• Experience building or working with large-scale distributed training and inference systems on Kubernetes.
• A proactive, self-driven attitude and willingness to work in a fast-paced, high-caliber team.

Compensation and Benefits

• Base salary: $150,000–$450,000 USD per year.
• Equity.
• Comprehensive medical, vision, and dental coverage.
• 401(k) retirement plan.
• Short- and long-term disability insurance.
• Life insurance.
• Various discounts and perks.

SpaceXAI is an equal opportunity employer.