Experience:

• Significant experience building and shipping complex AI or ML-heavy systems with ownership of architecture and production evolution
• Hands-on experience with agentic AI systems and frameworks such as ReAct-style agents, AutoGPT-like systems, LangChain, LangGraph, Semantic Kernel or similar
• Strong backend engineering fundamentals including advanced Python proficiency, API and service building, familiarity with databases, data pipelines, and cloud infrastructure
• Comfort working in ambiguous, fast-moving environments with loosely specified problems and expected ownership

Responsibilities:

• Define and build core AI agent architecture including reasoning, planning, execution, memory, and state management
• Build end-to-end AI Engineering Agents for autonomous experiment design, evaluation, analysis, and reporting
• Design and maintain complex harness components to support agent scalability for complex tasks
• Develop agent architectures combining reasoning, planning, tool orchestration, memory, and handling long-horizon execution with self-correction
• Integrate deeply with Lexsi backend, internal evaluation systems, data pipelines, alignment tooling, enterprise APIs, and proprietary data sources
• Design agents for enterprise and regulated environments ensuring inspectability, explainability, and defensibility of actions and outputs

Additional responsibilities:

• Own architecture and implementation decisions around planners, reasoning loops, memory models, and execution control
• Build robust tool orchestration and execution layers for reliable interaction with internal services, external APIs, and data systems with graceful failure recovery
• Embed alignment, safety, and interpretability into system design working closely with research teams
• Stress-test agent behavior in real-world conditions to identify edge cases, failure modes, and distribution shifts and iterate to improve reliability and correctness
• Contribute to design discussions, code reviews, and technical decision-making with strong ownership and accountability

Nice to have:

• Experience building long-horizon or stateful agent systems reasoning over time
• Prior exposure to AI alignment, interpretability, or safety tooling in production or enterprise settings
• Experience working in regulated or high-stakes domains requiring explainable and auditable system behavior
• Track record of debugging and improving systems with unpredictable real-world behavior

More skills:

AI engineering, agentic AI systems, ReAct-style agents, AutoGPT-like systems, API development, AI alignment, interpretability, safety tooling

Other:

• Flat organizational structure with high autonomy
• Strong bias toward engineers taking full ownership from architecture to production behavior
• Fast-moving environment valuing substance over polish and execution over rhetoric