Principal Engineer Agentic AI

Join Boston Scientific in Gurgaon, Haryana, India, at the forefront of AI-driven healthcare innovation. As a Principal Engineer Agentic AI, you will architect and deliver autonomous, goal-driven agents powered by large language models (LLMs) and multi-agent frameworks. This role involves designing and implementing agentic AI systems for reasoning, planning, and tool execution, leveraging frameworks like LangGraph, AutoGen, and CrewAI. You will develop essential components for context handling, memory, and API integration, enabling agents to interact with internal services and external tools reliably. Furthermore, you will create feedback loops and evaluation pipelines (e.g., LangSmith, RAGAS) to measure performance, factual grounding, safety, and latency. You will own backend services for scaling agent workloads, optimizing accelerator utilization, and managing costs. Responsibilities include embedding observability, drift monitoring, and alignment guardrails throughout the agent lifecycle. Collaboration with research, product, and security teams is key to translating emerging agentic patterns into production-ready capabilities. You will also mentor engineers on best practices for prompt engineering, tool-use chains, and agent deployment in regulated environments.

Key Responsibilities:

• Design and implement agentic AI systems leveraging LLMs for reasoning, multi-step planning, and tool execution.
• Evaluate and build upon multi-agent frameworks such as LangGraph, AutoGen, and CrewAI to coordinate distributed problem-solving agents.
• Develop context-handling, memory, and API-integration layers enabling agents to interact reliably with internal services and third-party tools.
• Create feedback-loop and evaluation pipelines (LangSmith, RAGAS, custom metrics) that measure factual grounding, safety, and latency.
• Own backend services that scale agent workloads, optimize GPU / accelerator utilization, and enforce cost governance.
• Embed observability, drift monitoring, and alignment guardrails throughout the agent lifecycle.
• Collaborate with research, product, and security teams to translate emerging agentic patterns into production-ready capabilities.
• Mentor engineers on prompt engineering, tool-use chains, and best practices for agent deployment in regulated environments.

Required Qualifications:

• 8+ years of software engineering experience, including 3+ years building AI/ML or NLP systems.
• Expertise in Python and modern LLM APIs (OpenAI, Anthropic, etc.), plus agentic orchestration frameworks (LangGraph, AutoGen, CrewAI, LangChain, LlamaIndex).
• Proven delivery of agentic systems or LLM-powered applications that invoke external APIs or tools.
• Deep knowledge of vector databases (Azure AI Search, Weaviate, Pinecone, FAISS, pgvector) and Retrieval-Augmented Generation (RAG) pipelines.
• Hands-on experience with LLMOps: CI/CD for fine-tuning, model versioning, performance monitoring, and drift detection.
• Strong background in cloud-native micro-services, security, and observability.

Requisition ID: 610421