Technical Architect - AI
Role details
Job location
Tech stack
Job description
Design the enterprise AI platform architecture spanning the LLM API gateway, GPU and compute allocation pools, sandbox provisioning, model registry, and security gate automation Define infrastructure standards, API gateway patterns, and reference architectures consumed by all AI delivery towers and partner integrations Establish guardrails for token metering, rate limiting, audit logging, DLP validation, SAST, DAST, dependency scanning, and model card review embedded in CI/CD Review security posture across all AI workloads with mapping to NIST AI RMF, AWS Well-Architected (including the Machine Learning Lens), and applicable enterprise compliance baselines Agentic AI and LLM Engineering Architect multi-agent systems using LangGraph, LangChain, and Model Context Protocol (MCP) for complex workflow orchestration, planning, and tool use Define patterns for ReAct, Chain-of-Thought, Tree-of-Thoughts, and agent-to-agent coordination across enterprise and customer-facing use cases Design and optimize Retrieval-Augmented Generation (RAG) systems, embedding strategies, and semantic search across structured and unstructured enterprise data Establish MLOps and AgentOps practices for deployment, evaluation, observability, and continuous improvement of agents and models in production AWS-Native Implementation Architect solutions on Amazon Bedrock, Amazon SageMaker, Amazon Q, Bedrock Agents, and Bedrock Knowledge Bases Define infrastructure patterns using Amazon EKS, AWS Lambda, ECS Fargate, API Gateway, EventBridge, SNS/SQS, Kinesis, S3, DynamoDB, Aurora, Redshift, Athena, OpenSearch, and Kendra Establish CloudFormation and AWS CDK templates and Terraform modules for isolated VPC sandboxes provisioned per project and per third-party partner Implement observability and FinOps using CloudWatch, AWS Cost Explorer, AWS Budgets, and chargeback reporting by team, project, and model Salesforce and SaaS AI Integration Define integration architecture with Salesforce Agentforce, Einstein, Data Cloud, and Service Cloud, including Apex, Flow, and Platform Event integration patterns with AWS-hosted agents and APIs Establish governance over enterprise SaaS AI licenses, including usage tracking, renewal governance, and redundancy elimination across business units Architect cross-system identity, authorization, and data exchange patterns spanning Salesforce, AWS, and partner endpoints Stakeholder and Delivery Leadership Partner with AIDO leadership, delivery tower leads, security, compliance, procurement, and program management to ensure platform adoption and consistent operating standards Produce enterprise-grade architecture artifacts, decision records, and operating model documentation suitable Mentor engineers across delivery towers and partner teams; lead architecture reviews and technical due diligence on partner-built systems
Requirements
Expert proficiency with LangGraph, LangChain, and agent orchestration frameworks Deep experience with Amazon Bedrock, SageMaker, and Amazon Q, including Bedrock Agents and Knowledge Bases Hands-on experience with Model Context Protocol (MCP), function calling, tool use, and structured output patterns Strong command of prompt engineering, evaluation harnesses, fine-tuning, and model optimization Working knowledge of transformer architectures, attention mechanisms, and multi-modal systems Machine Learning Classical ML (regression, tree-based ensembles, gradient boosting, clustering) and deep learning (CNNs, RNNs, transformers) across supervised, unsupervised, and reinforcement paradigms; feature engineering, hyperparameter optimization, cross-validation, drift detection, and model evaluation; end-to-end ML lifecycle on SageMaker spanning data preparation, training, deployment, monitoring, and retraining. AWS Platform SageMaker (Studio, Pipelines, Model Registry, Inference), Bedrock, EKS, Lambda, ECS Fargate, API Gateway, Step Functions S3, DynamoDB, Aurora, Redshift, Athena, OpenSearch, Kendra EventBridge, SNS/SQS, Kinesis, MSK CloudWatch, X-Ray, CloudTrail, AWS Config, GuardDuty, Macie, Security Hub IAM, KMS, PrivateLink, VPC design, and AWS Organizations governance Salesforce and Enterprise SaaS Salesforce Agentforce, Einstein, Data Cloud, Service Cloud, and Sales Cloud integration patterns Apex, Flow, Platform Events, and REST/Bulk API integration with external AI services Familiarity with enterprise identity providers, SSO, OAuth, and SCIM provisioning across SaaS estates Programming and Development Advanced Python with deep FastAPI experience for scalable, async API development Java proficiency sufficient to integrate with existing enterprise backend services Strong CI/CD background using AWS CodePipeline, CodeBuild, GitHub Actions, and Infrastructure as Code via Terraform and AWS CDK Containerization with Docker and orchestration with Kubernetes (EKS) Data and Vector Systems Vector store architectures using OpenSearch, Bedrock Knowledge Bases, Pinecone, Weaviate, or Chroma Embedding model selection, hybrid search, and reranking strategies Graph database experience (Amazon Neptune, Neo4j) for knowledge representation Data ingestion, masking, synthetic data generation, and DLP validation pipelines Experience Requirements 20+ years in software engineering with 5+ years focused on AI/ML systems 3+ years hands-on experience architecting and shipping production LLM and agentic AI applications Demonstrated success leading enterprise-scale AI platform builds with measurable business outcomes Track record architecting scalable cloud-native systems on AWS in regulated or large-enterprise environments Experience leading technical teams, mentoring engineers, and engaging executive stakeholders Education Bachelor's or Master's degree in Computer Science, AI/ML, or a related technical field AWS Certified Solutions Architect Professional or AWS Certified Machine Learning Specialty preferred Salesforce Certified AI Associate, AI Specialist, or Application Architect credentials is a plus