GenAI / Agentic AI Developer

Job Description We are looking for a hands-on GenAI / Agentic AI Developer to build LLM-powered applications, RAG solutions, and agentic AI workflows for enterprise use cases.

Key Responsibilities

• Build GenAI applications using LLMs, RAG, agents, and tool-calling workflows.
• Develop agentic solutions using LangChain, LangGraph, AutoGen, CrewAI, Semantic Kernel, or LlamaIndex.
• Design and implement multi-agent workflows such as planner, retriever, executor, validator, and human-in-the-loop agents.
• Build backend APIs using Python, FastAPI, Flask, REST APIs, and microservices.
• Integrate AI agents with enterprise systems, databases, APIs, document repositories, and cloud services.
• Implement document ingestion, embeddings, vector search, reranking, and retrieval pipelines.
• Deploy and monitor GenAI applications using Docker, Kubernetes, CI/CD, and cloud platforms.
• Support LLMOps including prompt/version management, model evaluation, monitoring, logging, and cost tracking.

Required Skills

• Strong hands-on experience in Python development.

• Experience with OpenAI, Azure OpenAI, AWS Bedrock, Anthropic Claude, Gemini, Llama, or Mistral.

• Hands-on experience with at least one agentic framework: LangGraph, LangChain, AutoGen, CrewAI, Semantic Kernel, or LlamaIndex.

• Good understanding of RAG, embeddings, vector databases, semantic search, and prompt engineering.

• Experience with vector stores such as OpenSearch, Pinecone, FAISS, Chroma, Weaviate, Milvus, Azure AI Search, or pgvector.

• Knowledge of REST APIs, cloud deployment, Docker, CI/CD, and software engineering best practices.

• Ability to work with structured and unstructured data including PDFs, documents, APIs, databases, and knowledge bases.

• Preferred Skills

• Experience with multi-agent orchestration, tool calling, memory, planning, reflection, and evaluation.

• Exposure to MCP, Graph RAG, Neo4j, knowledge graphs, or entity extraction.

• Knowledge of LLMOps tools such as LangSmith, MLflow, Phoenix, Ragas, TruLens, Arize, or OpenTelemetry.

• Experience with AWS Bedrock/SageMaker, Azure OpenAI/AI Search, or GCP Vertex AI.

• Understanding of AI guardrails, prompt injection prevention, PII masking, access control, and responsible AI.

• Must-Have

• Candidate should be able to clearly explain at least one end-to-end GenAI / Agentic AI project, including problem statement, architecture, tools used, deployment approach, evaluation method, and business impact.

• The base compensation range for this role in the posted location is 80,000 to 96,000

• Capgemini provides compensation range information in accordance with applicable nationa