Pytho AI · San Francisco
Build and deploy ML models for real-world defense operational problems, covering data exploration, feature engineering, evaluation, and production monitoring. Work includes data pipelines for text, imagery, geospatial, and sensor data, plus training and inference systems for edge deployment. Stack includes PyTorch, TensorFlow, Hugging Face, containerization and orchestration tools, with research into LLMs and multimodal systems.
Prepare first. Decide first. Win first.
At Pytho, we're on a mission to ensure warfighters are prepared for every challenge they face. We are committed to assembling a world-class team of exceptional problem solvers, combining expert engineering talent with deep experience in National Security.
We build AI for the most important problem in defense — making sure the force is ready to fight. Our software is deployed and in use with DoW partners today.
We're building out our AI Engineering Team. This is a chance to join at the earliest stage of a company where the engineering problems are genuinely hard — planning under uncertainty, multi-agent coordination, simulation, reinforcement learning — and the work is deployed in environments where the output actually matters.
You will work across the full ML lifecycle — transforming messy, real-world data into production-grade models deployed in high-stakes environments. You'll ship fast, work directly with the founders, and have outsized influence on the technical direction of the company.
AI/ML Development: Design, train, fine-tune, and deploy ML models to solve real-world operational problems. Own the lifecycle from data exploration and feature engineering to evaluation and production monitoring.
Data & Systems Engineering: Build scalable data pipelines for structured and unstructured data (text, imagery, geospatial, sensor data). Develop reliable training and inference systems optimized for performance and edge deployment.
Engineering Excellence: Integrate models into robust, scalable production systems with strong testing, observability, and CI/CD practices.
Research & Experimentation: Prototype and benchmark new modeling approaches (LLMs, multimodal systems) to improve performance, robustness, and mission impact.
Required
Preferred