What the work actually involves

You own the serving and infrastructure layer that AI-lab projects run on. Day to day that means provisioning GPU capacity on GCP (GKE node pools, Compute Engine, Cloud Run), keeping CUDA drivers and quotas sane, and deploying inference stacks with vLLM and/or Triton Inference Server behind containerized or serverless entry points. Alongside that sits pipeline work — Dataflow, Airflow or Cloud Composer moving data for training, evaluation, and inference workloads — plus the unglamorous half of the job: observability, autoscaling policies, reliability, and driving GPU cost per request down.

A note on the title: this listing appears as "Agentic Coding Annotator," but the description Turing published is a DevOps engineering role. Treat the requirements list as authoritative. If you were drawn here by annotation or evaluation work specifically, the adjacency is real — this infrastructure backs coding and agentic-behaviour evaluation for foundation-model customers — but the screen will ask you about node pools and inference throughput, not rubric design.

What the screen looks for

Turing's process is a profile review followed by an interview with the delivery team where needed, and the profile review is where most candidates are filtered. Screeners look for evidence that you have run GPU infrastructure in production rather than experimented with it: specific GCP services you have operated, a concrete vLLM or Triton deployment you tuned, a real incident you diagnosed. Expect follow-ups that go one layer deeper than your first answer — naming vLLM is cheap, explaining how you chose a `max_num_seqs` or KV-cache setting under a latency target is not.

Logistics

  • Fully remote, contractor engagement with no medical or paid leave.
  • 8 hours per day with at least 4 hours overlapping Pacific time — this is synchronous work, not async piecework.
  • Initial contract duration of 2 months, with an immediate expected start.
  • Pay is undisclosed on this listing; Turing typically negotiates an hourly contractor rate by region and seniority, and nothing here is guaranteed.