Software Platform Engineer II | Apply Now
Generate a McCoy IQ challenge in 30 seconds.
See how candidates think and approach the work this role demands, before the phone screen. We'll build a video challenge from this posting, and you can edit or share it before it goes live.
Key details
Why this role is notable
A 150-year-old insurer is standing up in-house agentic AI engineering on frontier models, signaling that agent orchestration and LLM operations are moving from pilots into core production systems at traditional financial services firms.
Job Description
Description and Requirements
- Design and implement production APIs and microservices in Python (FastAPI, async I/O).
- Build agentic AI workflows and the surrounding harness with LangGraph/LangChain — tool calling,
multi‑step orchestration, context & memory management, retries, and guardrails.
- Ship and operate production agents on frontier models (Claude, GPT, Gemini); optimize for latency,
cost, and reliability.
- Package, serve, version, and optimize ML/LLM models for low‑latency and batch/async inference
(MLflow/LangSmith or equivalent).
- Implement LLM features: embeddings, retrieval‑augmented generation (RAG), vector search, and prompt
& context engineering.
- Build resilient event‑driven integrations (Event Hubs/Kafka patterns), idempotent consumers, and
backfill/replay strategies.
- Containerize services (Docker), author manifests/Helm charts, and debug Kubernetes (AKS) workloads.
- Instrument services and models (logging, metrics, tracing), define SLOs, and lead incident triage
and remediation.
- Deliver automated tests and CI/CD for services and model pipelines; collaborate on release gating
and security scans.
- Partner with frontend (React/Next.js/TypeScript) to deliver end‑to‑end AI features.
Audit details(provenance, verification trail, raw fields)
Core fields
metlife:21396Provenance
www.metlifecareers.com|/en_US/ml/SearchJobsVerification trail
This posting hasn't been probed by our closure verifier yet. Stream C runs on a rolling schedule against postings approaching the close-decision threshold.
LLM enrichment
See how we measure for definitions, or our corrections log for known issues. Found something wrong? Flag a correction.