Production ML Engineer for LLM Systems & Guardrails

Latitude; Austin (TX); On-site; USD 224,000 - 279,000; Full time

Job summary

Fluidstack in Austin, TX is seeking a senior ML/LLM engineer to build and run models inside the company's operations, from framing to deployment and iteration in production.

You will own end-to-end models, ship agentic systems with guardrails and audits, and collaborate with data engineering and product teams to put predictions into the tools people use daily.

Experience shipping ML features to production and strong coding capabilities with AI tooling are required.

Qualifications

• Shipped ML or LLM features to production and owned them after launch.
• Built evaluation harnesses that tell the truth about model quality before users.
• Prefer the simplest model that works and can defend the choice.
• Hands-on with LLM APIs, fine-tuning, or retrieval systems on real business problems.
• Write production-grade code and work with AI coding tools.

Responsibilities

• Build ML/LLM systems that run inside Fluidstack operations: forecasting build timelines, flagging schedule risk, and extracting structure from vendor documents.
• Own models end to end, from problem framing and data through deployment, evaluation, and iteration in production.
• Ship agentic systems with guardrails, authorization, audit, and evals, so agents act on company systems.
• Partner with data engineering and product pods to put predictions in the tools people already use.

Skills

ML features shipped

LLM APIs experience

Production code

Evaluation harnesses

Retrieval systems

Job description

Fluidstack in Austin, TX is seeking a senior ML/LLM engineer to build and run models inside the company's operations, from framing to deployment and iteration in production.

You will own end-to-end models, ship agentic systems with guardrails and audits, and collaborate with data engineering and product teams to put predictions into the tools people use daily.

Experience shipping ML features to production and strong coding capabilities with AI tooling are required.