Machine Learning Engineer

Company Description

Decompute is a pioneering company enabling the fine-tuning of advanced AI models locally with minimal hardware requirements, such as 1 GB of VRAM. Our flagship platform, BlackBird, empowers users to create fully on-device AI agents, with applications in finance, legal, coding, voice, and image generation. Known for outperforming cloud-first solutions in complex document analysis, our proprietary models run efficiently offline on CPUs or GPUs, with a compact model size as small as 350M parameters. Decompute offers scalable solutions through both consumer and enterprise-grade offerings for on-premises or on-device AI needs.

Role Description

This is a full-time remote role for a Machine Learning Engineer. The selected candidate will design, develop, and implement machine learning models with a focus on on-device AI. Key responsibilities include developing innovative algorithms, training neural networks, evaluating model performance, and optimizing models for resource-constrained environments. Collaboration with cross-functional teams to integrate solutions into the BlackBird platform is also expected.

Qualifications

• Strong foundations in Pattern Recognition and Neural Networks
• Proficiency in Computer Science concepts, including Algorithms and efficient coding practices
• Solid understanding of Statistics and experience applying statistical methods to machine learning models
• Experience with optimizing machine learning models for both CPUs and GPUs
• Ability to work collaboratively in a remote setting and independently solve complex problems
• Masters or PhD in Computer Science, Artificial Intelligence, Machine Learning, or related fields is highly preferred
• Experience with on-device AI model fine-tuning or on-premises AI applications is a plus