Senior LLM engineer (TTS)

About Palabra** Palabra.ai is revolutionizing global communication with AI-powered real-time speech translation technology. Our mission is to give everyone the power to communicate without linguistic and cultural barriers. We are results-oriented and passionate about what we do. We’ve just raised $8.4 mln in a round led by the tier-1 US fund Seven Seven Six (776).

About the role Text-to-speech (TTS) is a crucial component in the speech-to-speech translation pipeline. We’ve developed our TTS model and trained it from scratch on a large corpus of data. It’s already on par with the best TTS models available on the market, but we want to improve its quality even further.

We seek an experienced ML/LLM engineer with solid hands-on experience in speech technologies and who is passionate about TTS.

In this role, you will:

• Develop algorithms for text-to-speech.
• Develop algorithms to improve zero-shot voice cloning between languages.
• Create pipelines to test the performance of trained models.
• Improve data processing pipelines.
• Analyze the current solution to achieve better performance for voice cloning and reduce streaming latency.
• Work closely with other members of the AI research team and the development team.

Who you are At Palabra.ai, we are an enthusiastic and globally spread team, each member fueled by a passion for AI-driven speech technologies. As part of our team, you'll be immersed in an environment that values initiative, mutual support, and personal accountability. You'll take full ownership of challenges from start to finish, readily acquiring any necessary knowledge to ensure good results, pursue excellence in every aspect of your work, and consistently exceed expectations.

Base qualifications

• BS/MS/PhD in Computer Science or a related field.
• 5+ years of industry experience as a ML or LLM engineer.
• 3+ years of experience in speech technology.
• Hands-on experience with LLMs.
• Understanding modern Machine Learning fundamentals, e.g., Transformers, GANs, diffusion networks.
• Proficiency in one or more neural network training frameworks, such as PyTorch or TensorFlow.
• Strong Python or C++ programming skills with an understanding of software development principles.

Will be a plus

• Hands-on experience with TTS, ASR or NLP.
• Experience with MLOps.
• Base knowledge of GPU architecture and extensive experience with GPU-related programming tools such as CUDA.
• Been a startup founder or an early-stage engineer.

What we offer

• Competitive salary.
• Opportunity to train LLMs on h100 GPU clusters.
• Stock Options: We want to share ownership and our collective success.
• Global Remote Work: Your location is secondary to who you are.
• Flexible Hours: We value flexibility as a remote-first company.
• Opportunity to collaborate with a passionate, experienced team committed to improving the world through innovative products.