Ai Engineer Jobs

About The Position

Requirements

• At least 3 years building and operating production software, including meaningful recent experience with LLM-powered systems.
• Experience shipping agentic systems that real users depend on and owning incidents when they fail.
• Fluency with agent technologies and practices, including tool calling, orchestration, context engineering, retrieval-augmented generation (RAG), evaluations, and tracing.
• Experience building and measuring retrieval systems over messy real-world corpora.
• Background in traditional machine learning, including supervised learning, feature engineering, and model evaluation.
• Treat evaluation as an engineering discipline.
• Ability to think in systems and customer outcomes, not only model metrics.
• Comfort with non-determinism and the ability to build reliable systems around it.
• Strong Python skills and comfort with TypeScript.
• Ability to ship quickly and verify results.
• Interest in large, complex, undocumented enterprise systems.

Nice To Haves

• Hands-on experience with enterprise platforms, APIs, and extension models, such as SAP, Salesforce, Workday, Oracle, Snowflake, MuleSoft, or ServiceNow.
• Experience with knowledge graphs, ontologies, or semantic models for messy real-world systems.
• Experience with code analysis, program transformation, or automated refactoring at scale.
• Experience with MCP, sub-agents, or agent skill and plugin architectures.
• Background in distributed systems or workflow engines, where partial failure is expected.
• Experience at an early-stage startup or as a founder.
• Familiarity with enterprise security and compliance, including SSO, RBAC, segregation of duties, PII handling, SOC 2, and data residency.

Responsibilities

• Design and ship production agents that understand enterprise landscapes, plan and execute changes across process, data, and code, and verify the results.
• Build typed, permissioned, well-documented tools that let agents interact with enterprise systems within human-approved authorization limits.
• Improve agent performance through prompting, context construction, tool-use strategies, and decision logic.
• Build retrieval systems for enterprise artifacts and metadata, including chunking and indexing, hybrid search, reranking, grounding, and evaluation.
• Manage agent context by deciding what information to provide, compress, or omit during long-running tasks.
• Apply classical machine learning to tasks such as routing, ranking, classification, anomaly detection, and confidence estimation when it is more appropriate than an LLM call.
• Build evaluation and monitoring systems, including task sets based on customer landscapes, regression coverage for deployments, and alerts for schema, authorization, or model-version changes.
• Instrument agent runs so model calls, tool invocations, decisions, and human approvals can be reconstructed for audit.
• Diagnose production failures using execution traces, identify root causes, and improve the system to prevent recurrence.
• Design guardrails, approval gates, and rollback paths for agents operating on live enterprise landscapes.
• Turn customer-specific patterns into platform capabilities that generalize across customers.
• Ship and verify software quickly, and focus on customer outcomes and reliable operation.