Senior Staff Engineer – Agentic Engineering

The Senior Staff Engineer is one of Marathon Health’s most senior individual contributors and serves as the engineering organization’s center of gravity for excellence in agent-assisted software development. The Senior Staff Engineer combines hands-on technical delivery with broad organizational influence, using coding agents and modern agentic engineering practices to ship better software, faster, and to raise the engineering bar across Marathon’s technology organization. This role partners with engineering and product leadership to shape both what gets built and how it gets built, drives architectural decisions that span team boundaries, defines reusable engineering patterns, and grows agentic and AI-native engineering fluency across Engineering, Data, and Infrastructure. The Senior Staff Engineer operates with significant autonomy on initiatives that cross organizational lines and translates the rapidly evolving practice of building software with agents into patterns and habits Marathon’s engineers can apply in daily work. 

ESSENTIAL DUTIES & RESPONSIBILITIES 

• Deliver cross-cutting engineering work — platform components, refactors, integrations, and high-leverage features — at a pace achievable only through disciplined use of coding agents (Claude Code, Codex, or comparable) and modern agentic engineering practices. 

• Set the technical standard for the engineering organization. Define and codify what good agentic engineering looks like at Marathon, including spec-driven development workflows, agent steering documents, custom Skills, MCP server patterns, subagent orchestration, and disciplined context engineering. Package these as reusable skills, plugins, templates, and internal documentation that other engineers can adopt without reinventing them. 

• Drive architectural decisions that span Engineering, Data, and Infrastructure. Partner with the Principal Architect and engineering leadership on systems design problems that cross team boundaries, lead ARB-level reviews, and bring strong fundamentals to API design, distributed systems, event-driven patterns, observability, and security within a HIPAA-regulated environment. 

• Build internal agents, MCP integrations, and RAG-backed applications where the return on investment is clear. Apply agent development frameworks pragmatically, with appropriate evaluation, guardrails, and cost discipline. 

• Influence both what gets built and how it gets built. Partner directly with engineering leaders, product partners, and internal stakeholders to shape initiative scoping, sequencing, and tradeoff decisions. Bring product judgment and technical discernment upstream into the question of whether a problem warrants engineering investment — not just downstream into how the work is delivered. 

• Mentor senior and staff-level engineers across the organization. Lead pairing sessions, design reviews, and brown-bags. Grow agentic and AI-native engineering fluency across Engineering, Data, and Infrastructure through hands-on collaboration and visible example. 

• Operate as a healthcare-aware engineer. Respect PHI handling, BAA requirements, audit trails, clinical safety considerations, and regulated change management, and help teams move efficiently within these requirements. 

QUALIFICATIONS 

• BS or MS in Computer Science or related field, or equivalent professional experience. 

• 10+ years of professional software engineering experience, with significant time at Senior, Staff, or Principal IC levels. 

• Deep systems design and architecture expertise, including distributed systems, API design, event-driven patterns, data flow, observability, and security. Agentic capabilities are leverage on top of strong fundamentals; we are not hiring for AI experience in lieu of engineering experience. 

• Hands-on experience across a modern engineering stack: cloud platforms (AWS, Azure, or GCP); strong proficiency in at least one of .NET, Python, or TypeScript and working fluency in the others; containers and Kubernetes (EKS, AKS, GKE) or comparable orchestrators; infrastructure-as-code (Terraform); CI/CD pipelines (GitHub Actions or comparable); modern observability tooling; and data platforms such as Snowflake and Kafka. 

• Daily-driver fluency with agentic engineering tooling. Routine use of a major coding agent (Claude Code, Codex, Cursor, or comparable) as a primary delivery vehicle, with direct experience managing context, orchestrating subagents, working with hooks, skills, and plugins, and recognizing the failure modes of long-horizon agentic tasks. Working knowledge of MCP — both consuming servers and authoring them — and of how Skills, slash commands, and agent steering documents fit together as a context-engineering toolkit. 

• Disciplined experience with spec-driven development applied to real production work. 

• Practical exposure to at least one agent development SDK or framework (such as OpenAI Agent SDK, LangGraph, or comparable), sufficient to make sound build, buy, or skip decisions about agent solutions. 

• Working knowledge of retrieval-augmented generation (RAG) patterns, including retrieval design, grounding, and the tradeoffs between RAG, fine-tuning, and longer-context approaches. 

• Demonstrated technical influence beyond individual contributions — patterns, libraries, frameworks, internal tooling, documentation, or mentorship that other engineers adopted and benefited from. 

• Comfort with product-adjacent work, including crisp problem framing, direct conversations with internal customers, MVP scoping, and confident decisions about whether a problem warrants engineering investment. 

• Strong written and verbal communication, with the ability to explain technical tradeoffs to executives, clinicians, and junior engineers in the appropriate register for each.

​ 

DESIRED ATTRIBUTES 

• Healthcare or other regulated-industry experience (HIPAA, HITRUST, SOC 2). 

• Experience with data orchestration and transformation tooling (Airflow, dbt, Databricks, or comparable) and modern data engineering patterns. 

• Experience evaluating LLM-powered systems, including eval harnesses, hallucination and regression testing, and cost and latency monitoring. 

• Track record of building and shipping internal frameworks, plugins, or skills that other engineers adopted at scale. 

• Cloud platform certification (AWS, Azure, or GCP) at the professional or architect level. 

• Discernment about what to build. Skeptical of agentic solutions when a simpler tool would do the job better, and able to recognize when an LLM call is the wrong answer. 

• Bias toward delivery. Prioritizes shipping production-ready solutions with a clear improvement path over pursuing exhaustive designs that fail to reach production. 

• Influence without authority. Changes how others work by being unambiguously strong at the work and generous with the lessons learned. 

• Comfort with ambiguity. Operates confidently where the established practices end and reports back with patterns the rest of the organization can use.