AI Engineer

Location: New York, USA;
Athens, Greece (On-site)

Twin Prime is an AI lab built for reasoning in the physical world, especially in defense and security settings. Founded by researchers and engineers from leading frontier model companies and quant funds, we build architectures that reason across complex sensor modalities in real time. We are headquartered in New York City and are currently expanding our footprint in the EU by way of Athens, Greece.

We are looking for AI engineers who combine a strong understanding of machine learning with the software engineering skills to turn research into reliable, deployable systems.

You will work closely with researchers and software engineers to implement new modeling approaches, build training and evaluation pipelines, and bring models into production. Ideal candidates have experience working across the model development lifecycle and are comfortable moving between experimentation, model optimization, and systems integration.

Areas of work include:

• Implementing and adapting models for video understanding, spatial reasoning, and multimodal learning.
• Building data processing, training, and evaluation pipelines.
• Designing experiments, analyzing model behavior, and translating findings into improvements.
• Optimizing inference performance, memory usage, and deployment on constrained hardware.
• Integrating models into reliable systems that process complex sensor data in real time.
• Building tools that make research workflows reproducible and accelerate the path from prototype to deployment.

The ideal candidate has hands-on experience developing machine learning models and strong software engineering fundamentals. Specifically:

• Experience training, fine-tuning, evaluating, and deploying deep learning models in research or production environments.
• Strong programming skills in Python and experience with PyTorch or a comparable deep learning framework.
• A solid understanding of deep learning fundamentals, including model architectures, optimization, and evaluation.
• Experience in computer vision, multimodal learning, generative modeling, or related areas.
• Ability to translate research papers and experimental ideas into working, maintainable code.
• Ability to diagnose problems across data, model behavior, and system performance.
• Comfort working closely with researchers and engineers on open-ended technical problems.

• Experience with distributed training or large-scale data pipelines.
• Experience with model compression, quantization, inference runtimes, or edge deployment.
• Familiarity with video, 3D, spatial, or other sensor modalities.
• Experience building performance-critical software in C++ or CUDA.