AI Ops Engineer Engineering · Los Angeles (Remote) · View role →

Furl, Inc.; Los Angeles, Northern (CA, KY); Hybrid; USD 140,000 - 200,000; Full time

Job summary

Furl, Inc. is seeking an experienced AIOps Engineer to bridge our agentic remediation platform with LLM-powered workflows in a remote Los Angeles setting. You’ll own the operational layer, ensure reliable inference, and drive cost-efficient performance at scale.

You’ll collaborate with AI and full‑stack teams, implement observation and evaluation pipelines, and optimize prompts, caching, and model selection to support rapid, secure remediation across platforms.

Qualifications

• Strong Python background with production services around LLM inference.
• Hands‑on experience with LLMOps or observability tooling (Arize, Phoenix, LangSmith).
• Comfortable deploying services in Kubernetes and cost optimization for LLMs.
• Experience with prompt management and caching strategies.

Responsibilities

• Own the operational layer between agentic workflows and LLMs; ensure reliable, observable, fast, and cost‑efficient inference at scale.
• Build and operate an LLM proxy layer with redundancy across model providers.
• Design prompt caching and semantic caching to reduce latency and cost while preserving quality.
• Run structured model testing and benchmarking across multiple providers to guide workflow power.
• Collaborate with AI and full‑stack engineers to improve observability, evaluation, and cost data.

Skills

Python

LLMOps

LLM Inference

API Orchestration

Gateways/Proxies

Observability

Arize

LangSmith

Kubernetes

Prompt caching

Semantic caching

OpenAI

Anthropic

Tools

Kubernetes

Job description

Engineering · Los Angeles (Remote) · Full-time

At Furl, we're building the first agentic remediation platform for IT security — technology that closes the gap between identifying risk and actually fixing it. Today, most organizations surface more vulnerabilities than they can handle. We're changing that by enabling safe, automated remediation across macOS, Windows, and soon Linux. We ingest security and operational signals, classify them into concrete remediation targets, and use an endpoint agent to collect real-time telemetry and execute fixes. Agentic workflows generate and validate fix plans, which are deployed safely at scale through policy-driven autonomy.

This is real infrastructure work with real-world impact. We're building systems that help teams focus on what truly matters — and we're doing it with care, clarity, and craftsmanship.

We're a real startup, we collaborate, build trust, and move fast together. It's a fun environment, but we won't sugarcoat it — the pace is demanding and sometimes uncomfortable. We value ownership over hierarchy and believe great work comes from people who are trusted to think, experiment, and take responsibility.

What You’ll Do

As an AIOps Engineer, you'll own the operational layer between Furl's agentic workflows and the LLMs that power them — keeping inference reliable, observable, fast, and cost‑efficient at scale.

• Build and operate an LLM proxy layer that provides redundancy and failover across model providers, so remediation workflows never depend on a single point of failure.
• Ensure every LLM message flows through our LLMOps platform for performance analysis and cost tuning, and integrate prompt management so services pull versioned prompts from Phoenix rather than hardcoding them.
• Design and implement prompt caching strategies to cut latency and inference spend without sacrificing output quality.
• Run structured model testing and evaluation — benchmarking candidate models against performance, cost, and accuracy targets to guide which models power each workflow.
• Collaborate with AI and fullstack engineers to make observability, evaluation, and cost data actionable across the platform.

Your Experience

• Strong Python background; experience building and operating production services around LLM inference (API orchestration, gateways/proxies, monitoring).
• Hands‑on experience with LLMOps or observability tooling such as Arize Phoenix, LangSmith, or similar (tracing, evals, prompt management).
• Comfortable deploying and operating services in Kubernetes.
• Experience with prompt caching, semantic caching, or inference cost optimization.
• Experience designing model evaluations and benchmarks across multiple LLM providers (OpenAI, Anthropic, open‑weight models).
• Experience in cybersecurity or IT B2B preferred.

Applicants must be authorized to work for any employer in the U.S. We are unable to sponsor an employment visa.

Applicants must be authorized to work for any employer in the U.S. We are unable to sponsor an employment visa.