- Build production RAG pipelines. over large, messy engineering document corpora (PDFs with tables, specifications, drawings, and scanned appendices), including layout- aware ingestion, structure-aware chunking, hybrid retrieval, and reranking.
- Engineer long-form structured generation. that produces multi-section engineering reports reliably, with deterministic document assembly, cross-section consistency, and numeric fidelity enforced through structured extraction rather than free-text generation.
- Implement source traceability and grounding. so every generated claim carries citations back to source documents, with grounding checks that flag unsupported content before it reaches a human reviewer.
- Design and operate agentic workflows. using Bedrock AgentCore and Step Functions, with bounded tool use, least-privilege per-session IAM, structured-output validation, and defense against prompt injection from untrusted document content.
- Build the evaluation harness. golden datasets, retrieval and faithfulness metrics, LLM- as-judge with human calibration, and regression suites that run in CI on every prompt or pipeline change.
- Harden PoC systems into production.



migrating sandbox prototypes onto scalable AWS services (managed vector stores, async job processing, container deployment) via CDK, with the platform teams coding and governance standards.
- Manage inference cost and latency. through prompt caching, batch inference, model routing by query complexity, quantization where self-hosting, and per-feature cost attribution.
- Integrate with the governed data layer. so that retrieval and querying respect Lake Formation permissions and per-user entitlements, never exposing data a user is not authorized to see.
- Keywords AWS Bedrock, Generative AI, RAG, LLM, Python, LangChain/LlamaIndex, Vector Databases (OpenSearch/Pinecone), AWS services (S3, Lambda, IAM, Step Functions), and production experience building AI/Agentic workflows.
- Mandatory Key Skills (atleast 1) AWS Bedrock, Generative AI, RAG, LLM, Python, LangChain/LlamaIndex, Vector Databases (OpenSearch/Pinecone), AWS services (S3, Lambda, IAM, Step Functions), and production experience building AI/Agentic workflows.
- Work Experience Required 5+ Year

Location - Mumbai, Delhi / NCR, Bengaluru , Kolkata, Chennai, Hyderabad, Ahmedabad, Pune,Remote

📌 AI Engineer - AWS (Mumbai)
🏢 Serendipity corporate services
📍 Mumbai