About the opportunity
Werker connects experienced practitioners with AI agents for paid conversations. In this mlops engineer opportunity, you will use your expertise to inspect a model deployment architecture, latency profile, or production incident. Sessions are focused reviews or walkthroughs. You can ask for the context needed to give a useful answer, and you remain responsible for staying within your professional scope.
Useful advice connects inference performance to capacity, rollback, data dependencies, and measurable service objectives. The aim is to help the other party make a better-informed next decision. Explain the evidence, the trade-offs, and what remains uncertain; where a question requires direct professional service, access to protected information, or another qualified decision-maker, say so clearly.
What you’ll do
- Review build, release, and rollback paths for model services.
- Identify serving bottlenecks, GPU-utilization issues, and cost drivers.
- Assess monitoring for latency, failures, drift, and data quality.
- Compare scaling strategies against operational complexity.
- Ask for missing context and distinguish verified information from assumptions before drawing a conclusion.
- Give a concise, practical recommendation with the trade-offs and limits made explicit.
What we’re looking for
- Production experience deploying and operating machine-learning systems.
- Knowledge of containers, cloud infrastructure, CI/CD, and monitoring.
- Ability to reason about throughput, latency, and reliability together.
- Comfort reviewing unfamiliar material and asking precise questions in a live conversation.
- Clear communication, professional discretion, and sound judgment about the limits of a short advisory session.
What a useful session produces
- A prioritized model-serving risk and performance assessment.
- A measurable deployment or monitoring improvement plan.
- A clear account of the assumptions and evidence needed before a higher-stakes decision is made.
How the engagement works
This is paid remote contractor work for US-based experts. After approval, your profile can be booked for focused voice conversations and, where useful, screen-based reviews or walkthroughs. Werker pays $120 USD / hour for completed sessions. You choose your availability and confirm each booking before the session.
Contract and payment terms
You would work with Werker as an independent contractor. Sessions are fully remote, and you set the times you can accept them. The length and scope of an engagement may change with project needs and performance; work can be extended, reduced, or ended early.
These opportunities do not require access to an employer’s, client’s, or institution’s confidential or proprietary information. Please do not share material you are not authorized to disclose. Werker pays weekly through Stripe or Wise for services completed.
At this time, Werker cannot support applicants who would rely on H-1B or STEM OPT authorization for this work.
Application process
- Submit your background and areas of expertise through the application form.
- Choose your availability and schedule a short AI interview about your work.
- Werker reviews the application before making a profile available for bookings.