Staff Machine Learning Engineer (AI)

Location: Sandy, Utah, USA; hybrid hirify.global-FLEX model with 2 days in the office and 3 days remote each week

Company

hirify.global develops AI, cloud, and digital software used by global businesses for customer experience, financial crime prevention, and public safety.

What you will do

• Evaluate and optimize AI models across Cognigy's agentic systems, including text-to-speech and speech-to-speech models.
• Track state-of-the-art ML models and assess their relevance to Cognigy use cases.
• Design model evaluations, benchmarks, human-judgment protocols, and automated metric validation.
• Apply fine-tuning, quantization, distillation, and efficient inference to improve quality, latency, and cost.
• Deploy and benchmark open-weight models on AWS, Azure, or GCP cloud platforms.
• Review teammates' optimization work and communicate recommendations to technical and non-technical stakeholders.

Requirements

• MS in computer science, machine learning, data science, or a related field.
• 3+ years of postgraduate hands-on experience training, fine-tuning, and evaluating ML models.
• Experience with model optimization and AI evaluation or benchmarking, including subjective or human-rated measures.
• Proficiency in Python and PyTorch or TensorFlow.
• Experience with cloud ML infrastructure for model testing and deployment.
• Ability to collaborate across functions, adapt to changing priorities, and present technical results clearly.

hirify.global-to-have">hirify.global to have

• Production experience evaluating or fine-tuning TTS, S2S, audio, or speech models.
• Exposure to agentic AI frameworks or conversational AI platforms.
• Experience with Docker, microservice deployment, and GPU inference serving.

Culture & Benefits

• Hybrid work through the hirify.global-FLEX model.
• Fast-paced, collaborative, and creative work environment.
• Opportunities for learning, growth, and internal career development across roles and locations.
• Work within a global organization operating across 30+ countries.