ZenteiQ has adopted Google Kubernetes Engine Inference Gateway for its synthetic scientific data-generation pipeline, lifting aggregate inference throughput by 4.07 times across its production clusters. The startup, spun out of the Indian Institute of Science and selected under the IndiaAI Mission, is building scientific foundation models designed to work with physical laws, governing equations and numerical solvers. Its pipeline creates prompt batches, generates candidate reasoning traces, scores responses and runs guardrail checks before verified datasets are passed into MaxText pre-training jobs, a process that had become a bottleneck because ZenteiQ operates with a fixed compute allocation. ZenteiQ had been running model pools on Kubernetes with standard ClusterIP round-robin load balancing across inference endpoints, which caused uneven engine queues and repeated prefix computation as it scaled work tied to its 9B- and 35B-parameter model schedules; prompts within a generation batch shared about 70% to 80% of their instructional context and task specifications, and throughput could vary by as much as half within a single run. ZenteiQ deployed GKE Inference Gateway, powered by the open-source llm-d distributed inference framework, in front of its vLLM accelerator pools, with the routing layer using live serving signals from vLLM instances instead of static rotation, monitoring queue depth and key-value cache use. ZenteiQ reported that aggregate throughput rose from 16,964 to 69,114 tokens per second per 1,000 accelerators, and