Top Tech Jobs & Startup Jobs

Research and develop autonomous multi-agent systems that plan, decompose tasks, choose experiments, evaluate outputs, and recover from failures. Design methods, coordination patterns, and evaluation frameworks for open-ended research workflows; run rigorous experiments, ship methods to production, and drive team research direction.

Write and optimize GPU kernels for training and inference, profile workloads with hardware counters, co-design kernels with researchers, integrate and benchmark kernels in training/serving stacks, and maintain kernel quality while sharing expertise across the team.

Design and run end-to-end ML experiments for autonomous research agents. Develop methods across RL, LLMs, multi-agent coordination, evaluate and scale results, and collaborate with engineers to productionize research.

Build, operate, and optimize production inference systems for large models. Own throughput, latency, cost-per-token, reliability, profiling, optimizations (quantization, custom kernels), observability, autoscaling, load testing, incident response, and postmortems.

Research and develop efficient inference methods (quantization, speculative decoding, sparse attention, MoE, distillation), run production-scale experiments, evaluate trade-offs, and partner with engineers to deploy improvements and publish results.

Lead post-training research including SFT, RLHF/RLAIF, reward modeling, and evaluation suite design. Run rigorous experiments, curate preference data, scale data and training pipelines, identify failure modes, and partner with data, infrastructure, and engineering to ship improvements into production.

Design, build, and operate training, evaluation, and deployment pipelines for ML research. Translate prototype research code into production-grade systems, own data pipelines, implement observability, set engineering standards, and collaborate closely with researchers to ensure scalable, reliable experiments.