About this role
The Senior AI Platform Engineer will design and own Arrowstreet Capital’s shared platform for firm-wide AI systems, including inference services, agentic platforms, developer tooling and observability. In this financial services environment, the platform must protect data, support end-to-end auditability and meet regulatory requirements. The role reports to the Head of AI Engineering and works with Security Engineering, AI integration and application teams, and infrastructure groups.
Responsibilities
- Build and operate managed LLM inference services, including Amazon Bedrock, model access management, versioning and routing across foundation models. Develop MCP servers, a registry and gateway, and authorization services connecting AI platforms to firm systems.
- Build usage, cost and adoption data pipelines and dashboards. Support AI-assisted developer tooling in Linux VDI environments, M365 and Excel integrations, and autonomous agent frameworks. Provide reusable inference and agentic platform components, including RAG, vector databases and model integration patterns.
- Partner with security and infrastructure teams on filesystem permissions, IAM policies, network allowlists, sandboxes and execution-time controls. Implement pre-execution guardrails and default-deny access; modifying or deleting production data requires specific authorization through a human-approval workflow. Keep inference traffic within the corporate network using VPC endpoints and PrivateLink, without public internet egress.
- Provide self-service onboarding, access controls, quotas, chargeback visibility and reference architectures. Operate centrally managed AI services and firm-wide applications.
Required qualifications
At least 10 years as an infrastructure, platform or systems engineer, including shared services used by multiple teams on-premises and on AWS. Strong expertise in AWS Bedrock, Azure OpenAI, and MCP registries, gateways, servers and authorization flows is required, along with production LLM workloads, platform-layer AI security in regulated or security-sensitive settings, and production agentic patterns involving tool use, function calling, RAG and human-in-the-loop workflows. Candidates must have built reliable, usable developer-facing platforms and communicate effectively across teams.
Preferred qualifications
Experience in regulated industries; M365 Copilot or Copilot Agents; observability pipelines using Splunk, ELK, Datadog or Grafana; containers and Kubernetes; model fine-tuning and ML lifecycle tooling; and DLP and data classification. The Boston-based, full-time position has a base salary range of $200,000–$325,000 per year, with annual discretionary bonuses and a benefits package.