Role Summary:

We are seeking a skilled and experienced AI Engineer to design, build, and operate production-grade generative AI capabilities. This role focuses on developing advanced Retrieval-Augmented Generation (RAG) pipelines, sophisticated multi-agent systems, and robust automated evaluation frameworks. The ideal candidate will bridge the gap between applied data science and rigorous software engineering, focusing on building scalable, secure, and cost-efficient AI-powered products.

Core Responsibilities:

- GenAI Application Architecture: Architect and deploy production-grade LLM applications using microservices (e.g., FastAPI) and robust software engineering practices, including APIs, integration testing, and CI/CD.

- Advanced RAG Engineering: Build and optimize end-to-end RAG pipelines, including document ingestion, semantic chunking strategies, metadata enrichment, vector database indexing (e.g., Azure AI Search), hybrid search (e.g., BM25 + vectors), and reranking to ground model outputs and minimize hallucinations.

- Agentic AI Workflows: Design and implement agentic AI solutions, incorporating tool-calling, state management, memory architectures, planning vs. reacting agent design, reflection loops, and multi-agent coordination using frameworks like LangGraph and AutoGen. Develop human-in-the-loop systems for verification and control.

- LLMOps & Lifecycle Management: Establish and manage operational standards for the AI lifecycle, including model fine-tuning, prompt versioning, semantic caching, rate limiting, and dynamic model routing to optimize for latency, cost (token economy), and performance. Implement CI/CD pipelines for AI/ML workloads.

- Model Evaluation & Observability: Develop and implement automated evaluation loops using "LLM-as-a-judge" methodologies to assess faithfulness, relevance, and toxicity. Monitor model drift, performance, and reliability using frameworks such as RAGAS, TruLens, and DeepEval.

- AI Safety & Governance: Implement strict guardrails (e.g., NeMo Guardrails, Llama Guard) to protect against prompt injection, data leakage, and other vulnerabilities, ensuring alignment with enterprise Responsible AI standards.

- Cross-functional Collaboration: Partner closely with data scientists, platform engineers, product owners, and business stakeholders to transition prototypes into stable, production-ready pipelines.

Required Skills & Experience:

- Programming Languages: Expert-level Python, SQL (Mandatory).

- LLM & GenAI Concepts: Deep understanding of tokenization, embeddings, prompt engineering, context windows, temperature/top-p tuning, and hallucination mitigation techniques. Experience with OpenAI/open-source LLM APIs, including structured outputs and function calling.

- GenAI Frameworks: Core: LangChain, LangGraph. Familiarity: LlamaIndex, CrewAI, AutoGen.

- Vector Databases: Experience with vector similarity search, metadata filtering, and optimization in databases such as Azure AI Search, PGVector, Pinecone, Qdrant, or Milvus.

- MLOps & Platform: MLflow (for model versioning, lineage, and tracking), Docker, Kubernetes. Experience with cloud platforms like Azure, Vertex AI (GCP), or AWS Bedrock/SageMaker. Proficiency with CI/CD automation and using AI coding assistants like GitHub Copilot.

- Evaluation & Guardrails: Experience with evaluation frameworks (e.g., RAGAS, TruLens, DeepEval, Arize/Phoenix).