Senior Software Engineer, AI Agentic Platform

Metropolis•Seattle, WA

2d•$177,000 - $216,000•Onsite

About The Position

Metropolis is seeking a Senior Software Engineer to architect and build its AI agent swarm, orchestration pipelines, and developer tooling. The role addresses distributed systems challenges at scale, makes build-versus-buy decisions for evolving AI infrastructure, and develops low-latency agent orchestration frameworks that support autonomous workflows for software development and business process automation. The engineer will collaborate with Application Engineering, Infrastructure, and Advanced Technologies to deliver real-world experiences. The position involves building production-grade agentic platform infrastructure; optimizing LLM inference routing, caching, and semantic search; establishing evaluation, tracing, and observability frameworks; and creating tools and automation that enable product teams to deploy AI agents safely. It also includes implementing security and operational guardrails, making architectural decisions, setting technical standards, evaluating AI tooling, and mentoring peers. Metropolis uses TypeScript, React, Scala, and Java; MySQL, PostgreSQL, and Snowflake; AWS; Git and GitHub; GitHub Copilot and Claude Code; and Datadog. Corporate team members follow an office-first model and are required to be on-site at least four days a week. The anticipated annual base salary is $177,000 to $216,000, with potential access or eligibility for healthcare benefits, a 401(k), disability coverage, basic life insurance, stock options, and bonus plans. Metropolis may use an automated employment decision tool in evaluating candidates and is an equal opportunity employer that provides reasonable accommodations during hiring.

Requirements

At least 6 years of backend software engineering experience focused on distributed systems, platform infrastructure, and high-load production environments.
At least 2 years of hands-on experience architecting and deploying agentic AI systems, LLM orchestration frameworks, or AI platform tooling.
Strong proficiency in Python or Java/TypeScript, with production expertise in asynchronous concurrency, API design, and system performance tuning.
Production experience with modern agentic orchestration frameworks and protocols such as LangGraph, Anthropic MCP, Google ADK, AutoGen, A2A, or LlamaIndex.
Expertise in microservices architecture, Kubernetes container orchestration, and AWS cloud platforms.
Observability and LLMOps expertise, including metrics, distributed tracing, structured logging, and debugging agent execution runs.
Knowledge of relational databases, vector databases, event streaming such as Kafka or SQS, and caching patterns.
Demonstrated experience using AI-powered development tools such as Claude Code, GitHub Copilot, or similar agentic coding tools for productivity and context engineering.
Excellent technical communication skills, including the ability to design and document complex AI architectures and collaborate across cross-functional teams.

Nice To Haves

Advanced degree in Computer Science, Machine Learning, or a related technical field.
Experience building customized agent harnesses, fine-tuning models, or implementing RLHF/RLAIF optimization loops.
Security compliance expertise involving data privacy, prompt injection defenses, secrets management, and access controls for financial or mobility data.
Contributions to open-source agent frameworks or MCP servers, or technical writing on agentic system architectures.

Responsibilities

Design and build production-grade agentic platform infrastructure, including harnesses, orchestration pipelines, tool-use abstractions, MCP integrations, and context and memory management systems.
Architect high-throughput, low-latency LLM inference routing, caching strategies, and semantic search interfaces to optimize cost, performance, and token usage.
Establish evaluation, tracing, and observability frameworks (LLMOps/AIOps) to measure agent accuracy, monitor execution paths, and debug non-deterministic workflows.
Build developer tooling, agent SDKs, and CI/CD automation that enable product engineering teams to deploy autonomous AI agents safely and efficiently.
Implement enterprise-grade guardrails, rate-limiting, sandboxing, and security policies for model execution and tool access.
Make architectural choices regarding service boundaries, data consistency, multi-agent communication protocols, and system reliability.
Drive technical standards, evaluate emerging AI tooling, and mentor engineering peers on agentic architectural patterns.
Collaborate across Application Engineering, Infrastructure, and Advanced Technologies.