AI Platform Engineer — Build Enterprise AI Infra

Zelis

Boston (MA)

Hybrid

USD 180,000 - 248,000

Full time

2 days ago

Be an early applicant

Get a reply from this employer

Your resume and cover letter, tailored to exactly what they’re hiring for.

Get past ATS filters

Benefits offered by this job

401k with employer match

Flexible PTO

Health, dental, vision

Job summary

Zelis is modernizing the healthcare financial experience and building Apollo, an internal AI platform. We’re seeking a hands-on AI Platform Engineer to influence the technical foundations other engineers will build on.

You’ll work at the intersection of AI, platform engineering, and enterprise delivery, shaping safe, scalable AI infrastructure used across Zelis. You’ll help translate evolving AI capabilities into dependable primitives, craft secure execution environments, and collaborate with

Qualifications

Strong software or platform engineering experience, including building and operating production systems used by other engineers.
Experience with distributed systems, AWS and cloud infrastructure, container orchestration, CI/CD, infrastructure as code, APIs, and observability.
Hands-on experience building with modern AI technologies such as foundation models, AI gateways, agent frameworks or harnesses, model evaluation, MCP, retrieval/context systems, or safe execution environments.
Strong programming skills in Python, TypeScript, Java, Go, or another modern language.
Practical understanding of model behavior, evaluation, prompt and tool design, token and latency tradeoffs, and model-provider integration.
Experience assessing new technologies with an eye toward security, reliability, maturity, cost, maintainability, and long-term platform value—not simply technical novelty.
Identity, secrets management, network boundaries, audit logging, and policy enforcement in enterprise environments.
Ability to work through ambiguity, run focused experiments, make clear technical decisions, and adjust course based on what you learn in production.
Strong communication skills and an interest in working directly with engineers, product partners, security, architecture, and senior stakeholders to drive adoption.
Experience building an internal developer platform, AI or agent platform, or engineering productivity product is especially relevant. Experience operating in healthcare, financial services, or another highly regulated environment is helpful but not required.

Responsibilities

Build the core agentic platform. Design and evolve capabilities across agent execution, workflow state and orchestration, context assembly, model routing, tool contracts, evaluation, observability, policy enforcement, and automated repair loops.
Create safe execution environments. Build and operate isolated agent runtimes across containerized, local, and simulation environments, supporting persistent sessions, testing, previews, adversarial review, and diagnose-and-repair workflows.
Develop reusable platform services. Help shape capabilities such as our AI Gateway, Context Engine, policy profiles, semantic models, project bootstrap tools, observability, and delivery integrations.
Make thoughtful technology bets. Evaluate emerging foundation models, coding agents, MCP technologies, workflow engines, developer tools, and commercial and open-source platforms. Run disciplined experiments and turn the right ones into supported Apollo capabilities.
Make security and governance part of the platform. Build identity‑aware access, isolation, policy-as-code, auditability, and controls for model and tool usage so engineers can move quickly without creating unnecessary risk.
Design for developers. Create clear APIs, interfaces, documentation, and paved paths that allow application engineers to take advantage of sophisticated AI infrastructure without needing to understand every element of the underlying control plane.
Learn from real‑world use. Work closely with forward‑deployed/agentic AI and business‑unit engineers, understand what is working and where they are getting stuck, and turn repeatable patterns into reusable platform capabilities.