YOUR ROLE
We are seeking a Senior Agentic AI Engineer to architect, build, and operate intelligent AI agents that execute complex, end-to-end business workflows.
This role goes beyond building conversational interfaces: you will develop reliable agentic systems that can reason over business context, plan and perform actions, collaborate with other agents, integrate with enterprise platforms, and involve people when judgment or approval is needed.
You will combine strong software engineering fundamentals with deep hands-on experience in LLM-based systems, agent orchestration, tool integration, and production operations. You will own technical outcomes from architecture and prototyping through deployment, evaluation, and continuous improvement, while helping establish engineering standards and mentoring other engineers.
WHAT YOU'LL DO
- Design and implement robust single-agent and multi-agent systems using patterns such as supervisor, planner, specialist agents, and delegated task execution.
- Build agent workflows that support planning, tool use, state transitions, retries, checkpoints, resumability, and human-in-the-loop approvals.
- Define clear boundaries between probabilistic agent reasoning and deterministic business rules, validations, and transaction execution.
- Develop reusable agent capabilities, tools, prompts, workflow components, and orchestration patterns that can be adopted across products.
- Manage agent context, memory, conversation state, task state, and long-running workflow lifecycle.
- Integrate and evaluate foundation models from providers such as OpenAI, Anthropic, Google, and AWS Bedrock.
- Engineer prompts, structured outputs, function/tool calling, context assembly, and model-routing strategies.
- Build retrieval-augmented generation (RAG) systems using embeddings, semantic search, metadata filtering, and vector databases.
- Improve quality and reliability through grounding, validation, confidence-aware workflows, fallback strategies, and hallucination mitigation.
- Optimize model selection, latency, throughput, token consumption, and inference cost without compromising business outcomes.
- Integrate agents with enterprise APIs, microservices, data platforms, and business applications.
- Build and consume Model Context Protocol (MCP) servers and tools using secure, well-defined contracts.
- Implement agent-to-agent (A2A) communication and interoperability patterns, including task lifecycle, status, input, and result handling.
- Design capability discovery, delegation, and invocation patterns for internal and external agents.
- Create resilient integration patterns for asynchronous operations, streaming, event delivery, and external system failures.
- Build scalable, fault-tolerant agent services using asynchronous and event-driven architectures.
- Implement durable workflow execution, persistence, checkpointing, idempotency, retries, timeouts, and recovery from partial failures.
- Establish observability across agent decisions, tool calls, traces, latency, token usage, costs, and business outcomes.
- Develop automated testing and evaluation strategies for agent behavior, tool correctness, regression detection, and end-to-end workflows.
- Implement enterprise security controls including authentication, authorization, tenant isolation, secrets management, auditability, and policy enforcement.
- Deploy and operate services using cloud-native infrastructure, CI/CD, containers, and appropriate runtime platforms.
- Lead technical design discussions and make pragmatic architecture decisions that balance speed, quality, scalability, and maintainability.
- Partner with product, platform, security, and domain teams to translate business problems into agentic solutions with measurable outcomes.
- Mentor engineers through design reviews, code reviews, pairing, and knowledge sharing.
- Evaluate emerging agent frameworks, protocols, models, and tools; recommend adoption based on evidence rather than novelty.
- Drive prototypes and proofs of concept into production-ready capabilities, including documentation and operational readiness.
- Other duties and responsibilities as assigned.