Senior AI Engineer

Job description

Job Description Summary We are seeking a highly skilled and innovative Senior Software Engineer – Agentic AI Engineer to join AIS Digital – Imaging360. The role will focus on designing, developing, and deploying production-grade AI solutions that leverage traditional AI/ML, Generative AI, Retrieval-Augmented Generation, knowledge graphs, and agentic workflows to solve complex healthcare imaging problems. The ideal candidate will bring 10+ years of software engineering experience, strong AI engineering depth, and a platform-agnostic mindset with the ability to evaluate and implement solutions across cloud, open-source, and enterprise AI ecosystems. Job Description Roles and Responsibilities In this role, you will: • Design, build, and deploy agentic AI systems that support autonomous reasoning, planning, tool use, multi-step execution, and human-in-the-loop workflows for Imaging360 use cases. • Develop scalable AI/ML and GenAI solutions across the full lifecycle, including data ingestion, feature engineering, model experimentation, evaluation, deployment, monitoring, and continuous improvement. • Architect and implement Retrieval-Augmented Generation solutions that combine structured and unstructured healthcare, product, operational, and workflow data with strong grounding, relevance, and traceability. • Design knowledge graph and GraphRAG solutions that model relationships across imaging assets, clinical entities, devices, and workflows, combining graph traversal with vector retrieval for multi-hop reasoning and explainable grounding. • Engineer evaluation and testing harnesses for agentic systems, including automated eval pipelines, golden datasets, LLM-as-judge scoring, regression suites for prompts and agents, simulation-based agent testing, and red-teaming for adversarial robustness. • Optimize token usage, latency, and inference cost through context engineering, prompt compression, prompt and semantic caching, model routing between frontier and small language models, batching, and structured output constraints. • Build platform-agnostic AI services using appropriate cloud-native, open-source, or enterprise AI capabilities while avoiding dependency on a single AI vendor or model provider. • Integrate LLMs, embedding models, orchestration frameworks, vector stores, APIs, and backend services into secure, reliable, and maintainable production applications. • Create rapid proofs of concept for emerging AI, GenAI, and agentic patterns; harden successful prototypes into reusable production components and engineering patterns. • Collaborate with product managers, architects, UX, data engineering, cybersecurity, quality, regulatory, cloud operations, and scrum teams to translate Imaging360 business needs into responsible AI solutions. • Define and implement AI evaluation methods for accuracy, relevance, grounding, robustness, latency, cost, safety, fairness, explainability, and operational reliability. • Drive engineering excellence through clean architecture, API design, automated testing, CI/CD, documentation, code reviews, design reviews, and production readiness practices. • Ensure AI solutions meet healthcare-grade expectations for privacy, security, auditability, data governance, compliance, and responsible AI adoption. Technical Skill Set • Programming and software engineering: Strong hands-on experience with Python and modern backend engineering; exposure to Java, TypeScript, or similar languages is preferred. • AI/ML and GenAI: Experience with machine learning, deep learning, NLP, LLM application development, embeddings, prompt engineering, fine-tuning or parameter-efficient tuning, and model evaluation. • Agentic AI: Experience designing AI agents, multi-agent workflows, tool orchestration, reasoning/planning loops, function calling, agent memory, guardrails, and workflow automation. • RAG and knowledge systems: Experience building retrieval pipelines, indexing strategies, chunking, reranking, metadata filtering, vector databases, hybrid search, and grounded response generation. • Knowledge graphs: Experience with graph databases (Neo4j, Amazon Neptune, or equivalent), ontology and entity modeling, entity resolution, GraphRAG patterns, and hybrid graph plus vector retrieval. • Agent evaluation and harness engineering: Experience with evaluation frameworks (Ragas, DeepEval, promptfoo, LangSmith, Langfuse, or equivalent), agent trajectory evaluation, offline and online eval loops, A/B testing, and CI-integrated evaluation gates. • Cost and performance optimization: Token accounting and budgeting, prompt caching, KV-cache-aware design, model right-sizing and routing, quantization and distillation awareness, and cost observability per request and workflow. • Interoperability protocols: Working knowledge of Model Context Protocol (MCP) for tool and data integration and agent-to-agent (A2A) communication patterns; experience building or consuming MCP servers is a plus. • Multimodal AI: Experience with vision-language models for medical imaging context, document understanding (reports, scanned forms), and multimodal RAG. • Architecture and integration: Strong understanding of APIs, microservices, event-driven systems, serverless or containerized architectures, distributed systems, and enterprise integration patterns. • Cloud and MLOps: Experience deploying AI services using cloud-native, container, or Kubernetes-based environments with CI/CD, model serving, observability, monitoring, and cost optimization. • Responsible AI and security: Knowledge of data privacy, secure AI patterns, access controls, content safety, hallucination mitigation, audit logging, governance, and healthcare compliance needs. • Platform-agnostic tooling: Working knowledge of one or more AI platforms or frameworks such as Azure OpenAI, Amazon Bedrock, Google Vertex AI, open-source LLMs, LangChain, LangGraph, LlamaIndex, Semantic Kern