Agentic AI and Data Engineer

Agentic AI and Data Engineer The Opportunity: As an experienced engineer, you know how to design, develop, and deliver production-grade agentic AI systems that demonstrate the practical value of generative AI, large language models (LLMs), and autonomous workflows. This role combines deep technical expertise with strong product skills to design AI applications that leverage prompting, retrieval-augmented generation (RAG), agentic orchestration, evaluation pipelines, and human-in-the-loop systems to deliver measurable impact. You will architect modular, reusable AI application patterns, integrate multiple model providers such as cloud-hosted, local, and hybrid, and apply modern GenAI stack capabilities, including structured prompting, tool use, workflow orchestration, and multi-modal reasoning. You will design solutions deployable across various contexts, from cloud-hosted platforms to portable, self-contained builds, optimizing for latency, cost efficiency, observability, and safety. You will rapidly prototype and iterate using AI-assisted development tools, validating hypotheses through eval-driven development and continuous experimentation. In this role, you’ll define the direction of mission-critical agentic systems by selecting and combining prompting strategies, RAG architectures, agentic workflows, and fine-tuned or foundation models as appropriate. You’ll be part of a large community of AI and ML engineers across the company, collaborating with data engineers, data scientists, solutions architects, and product owners to deliver world-class solutions. What You’ll Do: Design adaptable agentic AI architectures that support multiple model providers, tool ecosystems, modalities, and deployment modes. Build modular and reusable components for prompting, retrieval, orchestration, tool execution, memory management, and evaluation to enable rapid development of new AI capabilities. Integrate LLMs, embeddings, RAG pipelines, structured outputs, and long-context or memory mechanisms into production-ready systems. Apply advanced prompting techniques such as few-shot, chain-of-thought, tool-calling, and function-calling, orchestration frameworks such as LangChain or equivalent, and agentic architectures such as MCP, A2A, or similar patterns, to enable goal-directed autonomy with guardrails, observability, and human oversight, including planning, tool use, delegation, and recovery from failure. Design and implement evaluation frameworks, both offline and online, to measure correctness, robustness, safety, and business impact of AI systems. Optimize models and workflows for cost, latency, reliability, and scalability, using systematic benchmarking and experimentation. Develop data pipelines for ingestion, cleaning, chunking, embedding, indexing, and continuous refresh of structured and unstructured data for RAG and memory systems. Combine text, audio, vision, and other modalities in unified processing workflows, including document understanding, transcription, summarization, and cross-modal reasoning. Leverage vector databases, hybrid search, reranking, and retrieval optimization techniques to enhance grounding and reduce hallucination in RAG systems. Incorporate guardrails, safety filters, access controls, and monitoring mechanisms to ensure responsible and secure deployment of agentic AI systems. Deploy AI services securely and at scale on AWS or equivalent cloud platforms. Use containerizing, including in Docker or Kubernetes, or serverless approaches for flexible deployment. Apply CI/CD and eval-driven development best practices for AI systems, including automated testing of prompts and workflows, versioning of prompts and agents, and safe rollout of model updates. Use asynchronous programming and event-driven patterns to support scalable, long-running, or multi-agent workflows. Leverage modern build and packaging workflows to deliver optimized, portable application artifacts. Use AI assistance tools to accelerate development, deb ... (truncated, view full listing at source)