AI/ML Research Engineer

About UsAuric AI builds AI-first products exclusively for defense and intelligence applications. We're a self-funded, research-driven company working on unprecedented technical challenges that compete with the best in the world. Small team (10-15 people), 3-4 year runway, zero commercial pressure and pure focus on technical excellence.The RoleYou'll architect and build novel AI/ML systems for multi-modal intelligence analysis at scale. This involves designing new approaches for:- Cross-modal data fusion and correlation across heterogeneous sources - Geospatial-temporal reasoning and pattern discovery - Query understanding and semantic search for specialized domains - Anomaly detection in sparse, adversarial environments - Real-time inference at scale with strict latency requirementsThis is NOT:- Applying existing frameworks to new datasets - Fine-tuning foundation models - Standard RAG/retrieval systems - Incremental optimization workThis IS:- Designing novel architectures from first principles - Research that gets deployed to production within months - Direct user interaction with military/intelligence analysts - Building systems where failure has real-world consequencesWhat You Have1. Exceptional First Principles Thinking- You decompose complex problems to fundamentals and build novel solutions from scratch - You do not respect arbitrary convention2. Deep Algorithmic Foundations Beyond ML, you have mastery in:- Spatial algorithms: R-trees, k-d trees, geometric algorithms - Graph algorithms: temporal graphs, probabilistic graphical models - Temporal reasoning: causal inference, time-series analysis - Optimization: combinatorial optimization, approximate algorithms3. Novel Architecture Design You've built systems that don't fit existing frameworks:- Published novel architectures or deployed unprecedented systems - Created domain-specific neural network componentsWhat You GetTechnical Freedom:- No project managers, no arbitrary deadlines - You define what to build based on user needs - Full ownership: problem definition to deployment - Accept 70% failure rate in researchEnvironment:- Work with researchers from top institutions - Access to cutting-edge compute and specialized datasets - Direct access to end users for rapid iteration - Culture of intellectual honesty, zero politics

Apply Now