Teams at Pika

Design and build scalable backend and infrastructure for autonomous AI agents: real-time messaging, agent runtimes, LLM integrations, vector retrieval, APIs, databases, and production reliability while mentoring engineers and driving architectural decisions.

Lead inference acceleration and GPU-parallelism optimizations (TP, SP, PP); implement high-performance CUDA/NCCL kernels; deploy videogen and LLM models to production; collaborate with researchers, conduct code reviews, mentor engineers, and optionally improve training efficiency and resource utilization.

Contribute to large-scale pre-training and optimization of Videogen models, design and prototype high-throughput video/multimedia data pipelines, collaborate with researchers and engineers, and apply research techniques to production-scale generative video systems.