AI Engineer

Architect and maintain agent infrastructure, including orchestration logic, session handling, state management, and reliable backend service interfaces. Design robust tool integrations and MCP servers featuring precise boundary definitions and comprehensive error recovery handlers. Implement end-to-end tracing and observability for model calls, retrieval, and tool executions adhering to OpenTelemetry GenAI standards. Construct evaluation suites, benchmarks, and regression tests to validate model iterations and systematically capture production edge cases. Perform deep-dive root cause analysis on runtime failures, loops, context degradation, or hallucinated outputs to strengthen system resiliency. Enforce runtime guardrails, cost thresholds, RBAC permissions, output validation schemas, and human-in-the-loop validation triggers. Monitor industry advancements, emerging AI research, and architectural patterns, sharing technical insights across the engineering organization. Essential Deep software engineering expertise in a modern programming language, prioritizing algorithmic depth over superficial multi-language breadth. Strong computer science foundations in data structures, algorithms, memory allocation, and concurrent execution patterns. Practical experience with backend infrastructure: REST/gRPC APIs, relational/non-relational databases, caching tiers, message queues, and distributed systems resilience. Proficiency with Git, automated testing frameworks, CI/CD pipelines, and writing clean, maintainable software. Comprehensive mental model of LLM mechanics: context window constraints, tokenization, function calling protocols, non-deterministic behaviors, and hallucination management. Demonstrated history of delivering full-lifecycle technical projects across production environments, academic work, open source, or side initiatives. Methodical debugging skills with the ability to articulate systematic troubleshooting methodologies. Strong technical documentation and written communication skills to articulate complex distributed system behaviors. Expectations by Seniority Level Core requirements apply universally, with ownership and operational scale adjusting based on candidate seniority. Early Career: Focuses on technical foundations, algorithmic problem-solving, and curiosity. Prior agentic experience is not required; mentorship is provided to expand scope over well-defined modules. Mid-Level: Owns sub-systems end-to-end, translates ambiguous requirements into robust architecture, and demonstrates a proven track record of shipping and maintaining production services. Senior Level: Drives strategic architectural direction, possesses hands-on experience scaling non-deterministic systems, elevates team-wide engineering practices, and actively challenges technical assumptions. Nice to Have Hands-on experience developing application prototypes integrated with LLM APIs. Familiarity with Model Context Protocol (MCP) or custom tool integrations. Understanding of vector database indexing, embedding generation, and retrieval pipelines (RAG). Experience with cloud infrastructure (AWS preferred), containerization (Docker/Kubernetes), telemetry, or IaC tooling. Knowledge of GenAI security paradigms, prompt injection mitigation, and threat modeling. Active open-source contributions or community engagement. Out of Scope Extensive experience with specific agent frameworks, given the nascent nature of the ecosystem. Academic AI research background, specialized ML degrees, or mathematical derivation of backpropagation. Deep expertise in base model training or fine-tuning workflows. Flawless portfolio presentations; emphasis is placed on technical depth and problem-solving reasoning. Strictly conventional career trajectories; non-traditional and self-taught engineering backgrounds are evaluated equally. Technical Standards & Architecture Our engineering environment adheres to industry standards and operational frameworks, with continuous learning embedded into the engineering onboarding process. Model Context Protocol (MCP): Open standard managed via the Linux Foundation for integrating models with tool services and data stores. OpenTelemetry GenAI Conventions: Standardized instrumentation schemas for tracking model execution traces, step transitions, and tool invocations. OWASP GenAI & Agentic Security Guidelines: Threat modeling standards emphasizing containment strategies, least-privilege tooling, strict input sanitization, and authorization gates for critical actions. Eval-Driven Development: Mandatory evaluation validation prior to shipping behavioral changes, relying on internal domain-specific test harnesses rather than external benchmark metrics.

Desired Candidate Profile

What Are We Looking For Essential Deep software engineering expertise in a modern programming language, prioritizing algorithmic depth over superficial multi-language breadth. Strong computer science foundations in data structures, algorithms, memory allocation, and concurrent execution patterns. Practical experience with backend infrastructure: REST/gRPC APIs, relational/non-relational databases, caching tiers, message queues, and distributed systems resilience. Proficiency with Git, automated testing frameworks, CI/CD pipelines, and writing clean, maintainable software. Comprehensive mental model of LLM mechanics: context window constraints, tokenization, function calling protocols, non-deterministic behaviors, and hallucination management. Demonstrated history of delivering full-lifecycle technical projects across production environments, academic work, open source, or side initiatives. Methodical debugging skills with the ability to articulate systematic troubleshooting methodologies. Strong technical documentation and written communication skills to articulate complex distributed system behaviors. Expectations by Seniority Level Core requirements apply universally, with ownership and operational scale adjusting based on candidate seniority. Early Career: Focuses on technical foundations, algorithmic problem-solving, and curiosity. Prior agentic experience is not required; mentorship is provided to expand scope over well-defined modules. Mid-Level: Owns sub-systems end-to-end, translates ambiguous requirements into robust architecture, and demonstrates a proven track record of shipping and maintaining production services. Senior Level: Drives strategic architectural direction, possesses hands-on experience scaling non-deterministic systems, elevates team-wide engineering practices, and actively challenges technical assumptions. Nice to Have Hands-on experience developing application prototypes integrated with LLM APIs. Familiarity with Model Context Protocol (MCP) or custom tool integrations. Understanding of vector database indexing, embedding generation, and retrieval pipelines (RAG). Experience with cloud infrastructure (AWS preferred), containerization (Docker/Kubernetes), telemetry, or IaC tooling. Knowledge of GenAI security paradigms, prompt injection mitigation, and threat modeling. Active open-source contributions or community engagement.