AI Engineer - LLM Systems & Alignment

Vox AI is transforming the quick-service restaurant industry by pioneering voice-driven AI solutions tailored specifically for drive-thru automation and employee assistance. With a rapidly growing presence across multiple continents, we take pride in our pragmatic, innovative approach - leveraging AI to deliver seamless customer experiences and operational excellence. As we scale, we're seeking exceptional talent to join our Amsterdam-based team and help drive the next generation of voice technology.

We're looking for a passionate and experienced AI Engineer to specialize in Large Language Model (LLM) training and alignment. In this key role, you'll develop sophisticated training pipelines using reinforcement learning, supervised fine-tuning, and cutting-edge alignment techniques like DPO and ORPO. You'll create voice interaction systems that deliver natural, contextually-aware customer conversations and build robust API integrations enabling seamless interactions between our AI and restaurant systems. Your work will directly impact model performance, safety, and user satisfaction, positioning Vox AI at the forefront of conversational AI for the hospitality sector.

• Develop and optimize training pipelines incorporating reinforcement learning and supervised fine-tuning for LLM alignment


• Create and maintain voice interaction capabilities for conversational AI agents with natural language understanding


• Implement API integration frameworks allowing LLMs to interact with external systems and tools


• Build evaluation frameworks to measure model performance, alignment, and safety across different behaviors


• Develop inference optimization systems for low-latency model serving in production environments


• Create behavior-specific LoRA adapters for distinct use cases while maintaining a unified base model


• Implement monitoring systems for alignment drift detection in deployed agents

• Master's degree in Computer Science, Machine Learning, Artificial Intelligence or related field


• Demonstrated experience building and optimizing LLM training pipelines for large-scale models


• Proven expertise in alignment techniques including SFT, RLHF, DPO, and ORPO


• Strong experience with PEFT methods, particularly LoRA and QLoRA implementations


• Proficiency in developing and deploying multi-adapter architectures for different agent behaviors


• Experience with distributed training frameworks (DeepSpeed, FSDP, Megatron-LM)


• Knowledge of quantization techniques (FP8/INT8) for efficient model deployment


• Expertise in Python and deep learning frameworks such as PyTorch


• Experience with production ML systems and MLOps practices


• Knowledge of prompt engineering and instruction tuning methodologies

Preferred Qualifications:

• PhD in Computer Science, Machine Learning, or related field


• Experience developing multimodal models and systems combining text and audio modalities


• Knowledge of audio processing and voice-based AI systems


• Contributions to open-source LLM projects or research publications in NLP/ML


• Experience building commercial AI products with significant user adoption

• Venture-funded & growing fast – this is your chance to join early and make an impact.


• Build cutting-edge conversational AI systems with real-world impact


• Work with modern, open-source technology stack


• Hybrid work – Minimum 3 days/week in our Amsterdam office for high-impact collaboration.


• Equity included – we’re building something big, and we want you to grow with us.

If you thrive in dynamic environments, enjoy tackling complex challenges, and want to shape the future of voice AI technology with a global impact—this role at Vox AI is your opportunity. Apply now.