Senior ML Engineer

We are hiring experienced engineers at Physarum.ai to build enterprise-grade AI platforms on AWS, supporting hundreds of production AI agents and ML applications. The work spans AI platform architecture, agent development, and MLOps, including shared services that help teams build, deploy, and operate AI systems reliably, securely, and at scale.

Senior candidates are expected to own architecture decisions within their area of responsibility. For AI Architect and MLOps Engineer roles, platform architecture, governance, and production operations are core requirements.

Responsibilities:

• Design enterprise AI platform architecture on AWS, including runtime, integration, deployment, security, and operational patterns.
• Build Python-based platform services, APIs, and enterprise integrations.
• Develop agent runtime, workflow orchestration, session management, and tool-execution capabilities.
• Build agents, copilots, multi-agent workflows, RAG pipelines, and enterprise knowledge integrations.
• Integrate AWS AI services, including Amazon Bedrock and Amazon SageMaker, as appropriate to the workload.
• Build secure MCP servers and tool gateways.
• Implement model and agent evaluation, automated validation, tracing, and performance monitoring.
• Design scalable model-serving and inference infrastructure on AWS.
• Establish deployment patterns using Docker, ECS/Fargate, EKS, Lambda, CI/CD, and Terraform.
• Implement access controls, model governance, responsible AI, and enterprise compliance requirements.
• Own reliability, scalability, observability, and AI cost optimisation.

Requirements:

• Experience in applied AI, backend engineering, ML platforms, MLOps, or distributed systems.
• Strong Python, API design, object-oriented programming, and asynchronous programming skills, with FastAPI or a similar framework.
• Hands-on experience building agentic systems, LLM applications, RAG pipelines, or tool-calling workflows.
• Strong AWS experience covering AI services, compute, storage, identity, networking, and monitoring.
• Practical familiarity with relevant AWS services such as Amazon Bedrock, ECS/Fargate, EKS, Lambda, S3, DynamoDB, API Gateway, IAM, VPC, Secrets Manager, and CloudWatch.
• Experience with agent frameworks such as LangGraph, LangChain, or equivalent.
• Experience with Docker, CI/CD, and Terraform.
• Understanding of event-driven systems, orchestration, retries, state management, and production reliability.
• Ability to translate enterprise requirements into architecture and implementation decisions, explain trade-offs, and guide engineering delivery.

For AI Architect roles:

• Experience designing enterprise AI platforms, agentic architectures, and multi-agent systems on AWS.
• Strong knowledge of AWS architecture, Kubernetes/EKS, platform engineering, APIs, and MCP-based integrations.
• Ability to define security, governance, responsible AI, and compliance controls, including IAM, network isolation, encryption, and auditability.
• Ownership of architecture standards, technical reviews, observability, AI FinOps, and operational readiness.
• Ability to select appropriate AWS managed services and deployment patterns based on reliability, scalability, cost, and operational requirements.

For MLOps Engineer roles:

• Experience designing and building MLOps platforms on AWS across model validation, governance, deployment, and monitoring.
• Ability to architect scalable model-serving and inference systems using SageMaker or container-based infrastructure on EKS/ECS.
• Experience building CI/CD pipelines and automated testing and evaluation frameworks for ML workloads.
• Familiarity with model registries, experiment tracking, reproducible pipelines, and tools such as MLflow or Kubeflow.
• Strong understanding of container orchestration, Terraform, reliability, capacity planning, and cost management.

Additional useful experience:

• Amazon Bedrock AgentCore or comparable agent infrastructure.
• Workflow engines such as AWS Step Functions or Temporal.
• Event-driven processing using SQS, SNS, and EventBridge.
• Evaluation and tracing tools such as Langfuse, LangSmith, or OpenTelemetry.
• Knowledge graphs, vector databases, and enterprise data platforms.
• Model routing, prompt and response caching, and retrieval optimisation.
• GitLab CI/CD and reusable deployment templates for AWS workloads.