This position is listed on behalf of a partner company, who manages all applications and next steps. Our partner is looking for a Senior Artificial Intelligence/Machine Learning Engineer based in India. The Senior Artificial Intelligence/Machine Learning Engineer will design and deliver production-grade AI systems that transform complex, unstructured documents into accurate, structured data. This hands-on engineering role focuses on agentic AI, document intelligence, large language models (LLMs), and retrieval-augmented generation (RAG) within scalable AWS environments. You will build intelligent extraction agents, resilient document ingestion pipelines, and confidence-based routing systems that balance automation with human oversight. The role emphasizes reliable outputs, secure domain-specific knowledge, measurable model performance, and responsible AI governance. You will also optimize inference costs and latency while developing evaluation frameworks and feedback loops that continuously improve system quality. Working within a collaborative, cross-functional engineering team, you will mentor other engineers and help establish reusable technical patterns for dependable, high-impact AI solutions. Accountabilities - Agent development: Design, build, test, and maintain production-grade extraction agents that convert unstructured document content into structured target schemas while supporting multiple schema variants. - Document ingestion engineering: Develop robust ingestion pipelines capable of processing degraded scans, multi-generation photocopies, mixed image-and-text PDFs, and documents with inconsistent layouts. - Confidence modeling and routing: Build field-level confidence models and routing mechanisms that allow high-confidence outputs to proceed automatically while directing uncertain results to human review. - Knowledge engineering and RAG: Build and configure retrieval-augmented generation knowledge bases that provide agents with secure, domain-specific context, improve grounding, and reduce hallucinations when incorrect values could have significant downstream consequences.
- Human-in-the-loop governance: Implement review and escalation workflows, confidence thresholds, and output controls. Ensure reviewer corrections are captured and used as feedback signals to improve future performance. - Change detection and reprocessing: Develop comparison workflows that evaluate new extraction results against existing records and highlight meaningful changes. - Model and pipeline optimization: Configure model selection, prompt and context strategies, token usage, inference costs, and latency across the end-to-end AI pipeline. - Evaluation and monitoring: Establish mechanisms for assessing extraction accuracy, monitoring production performance, identifying quality issues, and improving system reliability over time. - Technical ownership: Own the AI system lifecycle from document ingestion and inference through evaluation, monitoring, governance, and feedback-driven improvement. - Engineering mentorship: Provide technical guidance to mid-level AI engineers and establish standardized patterns for extraction, confidence estimation, agent orchestration, and system design. - Cross-functional collaboration: Work with relevant engineering and product stakeholders to translate business needs into dependable AI capabilities and scalable technical solutions. Requirements: - Deep experience in AWS architecture, including hands-on work with Amazon Bedrock in production environments. - Strong expertise in model selection, prompt engineering, context engineering, model evaluation, and optimizing inference costs and latency. - Substantial experience with document AI and optical character recognition (OCR), including processing degraded inputs using Amazon Textract or equivalent technologies.