Fullstack AI Engineer-Agentic

Key responsibilities

• Build and ship LLM-powered agent features tool calling, structured outputs, retries and fallbacks — within established architectural patterns.
• Implement retrieval and structured data access to support agent reasoning.
• Write tests for non-deterministic behaviour, including evaluation cases and regression checks.
• Instrument features for observability — tracing, token usage and error reporting.
• Own assigned features end to end, from implementation through hardening and release readiness.
• Integrate agent capability with surrounding platform services and APIs.
• Use AI coding assistants within defined engineering and security guardrails: independently validate generated code, verify it through unit, integration and regression tests, and never treat generated output as production-ready without human review.
• Take part in code review and act on feedback from senior engineers and technical leads.

Required skills and experience

• Strong Python, including async patterns and building services that call external APIs reliably.
• Demonstrable production experience with LLM APIs — not prototypes or notebook work alone.
• Practical experience with tool and function calling, structured outputs, retries and fallback handling.
• Retrieval-augmented generation or structured retrieval over a document or data corpus.
• Testing discipline for probabilistic systems: evaluation cases, regression checks, and a way to tell whether a change improved or degraded behaviour.
• Working full stack capability — able to build or extend the APIs and interfaces around the features you deliver.
• Disciplined use of AI coding assistants within defined engineering and security guardrails: you understand and validate generated code before accepting it, check it for maintainability, architecture alignment, security, privacy, licensing and performance, and can explain and defend anything you submit whether or not an assistant helped produce it.
• Comfortable working in an existing codebase and with ambiguity in agent behaviour.

Preferred qualifications

• Exposure to an agent framework such as LangGraph, LlamaIndex, CrewAI or AutoGen.
• Cloud platform experience (AWS or Azure) and containerised deployment.
• QA or test-automation domain knowledge.
• Cost and token-efficiency awareness for LLM workloads.