GenAI / AI-ML Engineer

Role summary We are seeking an experienced GenAI / AI-ML Engineer to design, build, deploy and operate production-grade AI/ML and Generative AI solutions on AWS. The role requires strong Python engineering, solid understanding of conventional machine-learning algorithms, practical experience with LLMs, RAG and agentic AI, and strong hands-on expertise with AWS AI/ML and Generative AI services. The engineer will be responsible for taking AI solutions from problem definition and solution design through development, evaluation, production deployment and optimization. The candidate should be capable of making architecture and technology decisions based on scalability, latency, reliability, security, model quality and cost. The role also requires a strong understanding of LLM token consumption, model pricing and AWS service/infrastructure costing, with the ability to design solutions that balance business requirements, technical performance and operational cost. Key responsibilities • Translate business problems into measurable AI/ML and Generative AI objectives, solution approaches and evaluation criteria. • Design and implement production-grade LLM, RAG, conversational AI and agentic AI solutions on AWS. • Build RAG pipelines including document ingestion, chunking, embeddings, metadata filtering, vector retrieval, reranking and generation. • Design and implement Amazon Bedrock Agents and Bedrock AgentCore-based agentic solutions, including tool use, memory/state, runtime execution and orchestration as applicable. • Work with Amazon Bedrock foundation models and evaluate models based on quality, latency, token consumption and cost. • Build integrations between LLMs/agents and enterprise APIs, databases, applications, tools and knowledge sources. • Develop and evaluate supervised, unsupervised and deep-learning models based on business requirements. • Apply appropriate machine-learning algorithms for classification, regression, clustering, anomaly detection and other relevant use cases. • Perform data preprocessing, feature engineering, model training, validation, hyperparameter tuning and error analysis. • Use Amazon SageMaker for appropriate ML development, training, experimentation, deployment and model lifecycle requirements. • Establish evaluation mechanisms for LLM response quality, retrieval quality, groundedness, hallucination, latency and cost. • Analyze and optimize LLM input/output token consumption, context size and model selection to control GenAI costs. • Estimate and optimize AWS infrastructure and service costs across model inference, compute, storage, APIs, vector search and other components. • Design end-to-end AWS architectures using appropriate managed AI/ML, compute, data, integration and security services. • Implement security, access control, logging, monitoring, observability and operational controls for production AI systems. • Deploy and operate AI/ML solutions on AWS and troubleshoot latency, scalability, reliability, model quality, infrastructure and cost issues. • Monitor deployed ML models for performance, drift, reliability and business-aligned metrics. • Collaborate with data engineers, software engineers, architects and product teams to build reliable production solutions. • Produce reusable components, technical documentation, architecture decisions and implementation guidance for delivery teams. Required qualifications and experience • 5+ years of relevant experience in AI/ML engineering, machine learning, Generative AI, applied data science or software engineering. • Strong hands-on Python development experience with good understanding of object-oriented design, API integration, debugging and production engineering. • Demonstrable experience taking AI/ML or GenAI solutions from POC/experimentation through production deployment. • Mandatory hands-on AWS experience implementing and deploying AI/ML or GenAI solutions. • Strong hands-on experience with Amazon Bedrock and foundation models. • Hands-on experience with Amazon Bedrock AgentCore and/or Bedrock Agents for agentic AI solutions. • Hands-on experience with Amazon SageMaker for ML model development, training, deployment or lifecycle management. • Experience with AWS services supporting GenAI architectures, including Amazon S3, Amazon OpenSearch Service/Serverless, AWS Lambda, API Gateway, Step Functions, CloudWatch, IAM, KMS and Secrets Manager. • Strong understanding of LLMs, embeddings, vector search, prompt/context engineering and RAG. • Working experience with agentic AI, tool/function calling, agent orchestration and state management. • Strong understanding of conventional machine-learning algorithms, model evaluation and statistical reasoning. • Ability to select appropriate algorithms, models and AWS services based on the business problem. • Strong understanding of trade-offs between model quality, accuracy, latency, scalability, reliability and cost. • Understanding of LLM token economics, including input/output tokens, context consumption and model pricing. • Ability to understand and estimate AWS service/infrastructure costs for an AI/ML solution. • Ability to translate business requirements into secure, scalable, maintainable and cost-effective AWS AI/ML architectures. Mandatory Skills Python, Pandas, NumPy, Scikit-learn Conventional ML – Classification, Regression, Clustering, Anomaly Detection Feature Engineering, Model Training & Evaluation Cross-validation, Hyperparameter Tuning, Error Analysis AWS – Bedrock, SageMaker, Knowledge Bases, OpenSearch/OpenSearch Serverless RAG, Embeddings, Vector Retrieval Agentic Workflows, Tool/Function Calling, State Management LLM Evaluation, Groundedness, Hallucination Mitigation LLM Token Consumption & Cost Estimation AWS Service/Infrastructure Cost Estimation & Optimization Good-to-Have Skills Docker, CI/CD, Kubernetes/EKS MLflow / SageMaker MLflow PyTorch/TensorFlow Advanced MLOps & Model Monitoring Fine-tuning, PEFT, LoRA, QLoRA Advanced NLP, Computer Vision, Forecasting Bedrock Agents / AgentCore Multimodal GenAI Voice AI / Conversational AI GraphRAG / Knowledge Graphs Bedrock Guardrails / Evaluation MCP / A2A & Agent Interoperability Advanced LLM Observability & Evaluation Advanced Multi-Agent Architectures