Risk, policy and third party. Run the information security risk register as a decision-making tool, own the policy lifecycle and exception register, and assess the vendors and partners we integrate with.
Incident response and regulatory notification. Own breach assessment and the notification decision across the jurisdictions we operate in, alongside Legal. In healthcare this is the highest consequence judgement in the role.
Finding what is broken before someone else does. Go looking. Read the infrastructure code, pull the access review output, check that the alert a policy promises is actually configured. When you find a gap, bring it quantified, costed and sequenced.
...
Lead the design, specification, and engineering of mechanical building services (HVAC, fire protection, and other M&E systems). Develop detailed design drawings, schematics, and technical documentation.
Prepare and review mechanical datasheets, specifications, and technical requisitions for packages and rotating equipment.
Evaluate vendor bids, perform technical bid clarifications, and recommend vendor selection based on compliance and lifecycle cost.
...
You will work across the entire stack — from frontend experiences and APIs to distributed systems, data infrastructure, LLM orchestration, agentic workflows, evaluation systems, and autonomous optimization loops.
The ideal candidate has deep hands-on experience with Claude/Anthropic models, agentic architectures, tool-use, multi-step reasoning, autonomous execution loops, and AI-driven self-optimization.
You should be comfortable asking: "How can we make the system improve itself rather than requiring an engineer to manually optimize every workflow?"
...
You will work across the entire stack — from frontend experiences and APIs to distributed systems, data infrastructure, LLM orchestration, agentic workflows, evaluation systems, and autonomous optimization loops.
The ideal candidate has deep hands-on experience with Claude/Anthropic models, agentic architectures, tool-use, multi-step reasoning, autonomous execution loops, and AI-driven self-optimization.
You should be comfortable asking: "How can we make the system improve itself rather than requiring an engineer to manually optimize every workflow?"
...
Full Stack EngineeringFrontend: React, Next.js, TypeScript, modern component architectures, state management, real-time and streaming AI interfaces, agent activity and execution interfaces, data visualization.Backend: Node.js, TypeScript, Python, REST APIs, GraphQL, WebSockets and streaming, event-driven architectures, background workers, job queues, distributed systems, authentication and authorization.
Distributed SystemsDesign systems that reliably execute thousands or millions of AI and data-processing tasks. Kubernetes, Docker, Cloud Run and serverless, message queues, Redis, Kafka or equivalent, distributed job processing, concurrency management, rate limiting, retries, idempotency, fault tolerance, observability. You know how to build systems that stay reliable when agents fail, APIs time out, models hallucinate, or downstream services go away.
Data & Learning InfrastructureBuild the infrastructure agents need to learn from historical executions. PostgreSQL, BigQuery or equivalent data warehouses, ClickHouse or analytical databases, vector databases, embeddings, retrieval systems, event logs, feature stores, analytics pipelines, data ingestion.
...