Member of Technical Staff

About the role

SpaceXAI is building AI systems designed to interpret the universe and support human knowledge acquisition. You will join a flat, high-output team focused on developing the Grok Voice Model to provide natural, low-latency, and multilingual spoken interactions. In this capacity, you will own the end to end lifecycle of voice model development from data strategy to deployed service. The role demands comfort with ambiguity and a bias for rapid, iterative experimentation in a fast moving environment. You will partner closely with product and research to translate ambitious voice interaction concepts into reliable engineering solutions. Success in this position is defined by your ability to ship high quality models that balance conversational fluency with strict factual grounding. You will be expected to independently drive technical decisions while communicating clearly across a distributed, high performing team.

Key facts

What you'll do

• Develop and manage pipelines for large scale speech data, including audio collection, synthetic data generation, and automated annotation.
• Execute pre training and post training for speech language models using supervised fine tuning and reinforcement learning to ensure conversational fluency and factual accuracy.
• Create evaluation frameworks to track latency, model quality, human preference, and factual grounding.
• Collaborate with product teams to deploy voice models into real time applications and manage the lifecycle from prototype to global scale.
• Design and implement scalable data ingestion and transformation workflows that handle diverse audio sources and streaming formats.
• Conduct rigorous experimentation to optimize model performance, focusing on multilingual support and natural spoken style.
• Instrument production systems to monitor model behavior, data quality, and serving metrics in real time.
• Partner with research and product stakeholders to define success criteria and iterate on voice capabilities based on user feedback.
• Build and maintain reusable tooling that standardizes evaluation, logging, and deployment processes for voice experiments.
• Act as a technical owner for reliability and performance, ensuring that voice services meet strict latency and availability targets.
• Drive the creation of synthetic data strategies to augment training sets and improve robustness to varied speaking conditions.
• Lead post mortems on model regressions or serving incidents to identify root causes and preventive actions.
• Contribute to open source components where appropriate to accelerate development and share best practices.
• Mentor junior engineers on scalable machine learning practices and the nuances of speech technology.

Requirements

• Expert level Python programming skills for AI and machine learning systems.
• Practical experience with Spark and Ray for processing and augmenting large scale datasets.
• Proficiency in JAX or PyTorch for training speech language models with a focus on multilingual performance and natural spoken style.
• Ability to build evaluation infrastructure, including A B testing, human preference studies, and objective metric tracking.
• Experience with distributed training and inference systems running on Kubernetes.
• Strong understanding of speech processing concepts, including data preparation, feature extraction, and model evaluation for audio modalities.
• Demonstrated ability to work effectively in a fast paced, high growth environment with evolving priorities.
• Excellent written and verbal communication skills to coordinate with global teams and articulate technical trade offs.
• Commitment to excellence through rigorous experimentation, code review, and documentation.
• Willingness to dive into complex, poorly defined problems and deliver pragmatic solutions under tight timelines.

Nice to have

• Preferred items are not specified for this role.

Practical notes

Salary range: $150,000 - $450,000 USD. Total compensation includes equity, medical, vision, and dental insurance, 401(k) plan, disability and life insurance, and additional company perks. SpaceXAI is an equal opportunity employer.

The role is full time, and candidates must be eligible to work in the country where the position is located without requiring sponsorship at the time of hire. No relocation support is provided for this position. Applicants must meet all listed requirements; partial substitutions of skills or experience will not be accepted. The successful candidate will be required to pass standard background checks and security screenings as applicable to the role. Start dates will be coordinated based on team availability and candidate readiness. All hiring materials, assessments, and interviews are designed to evaluate technical capability and alignment with SpaceXAI standards.