与之科技 builds infrastructure for deploying and optimizing large-scale deep learning models, with a tech stack centered on PyTorch, TensorRT, and ARM—a combination indicating focus on edge inference and model compression. The company is actively working on automated pruning and quantization pipelines alongside large-scale model training, but hiring has stalled (zero roles posted in the last 30 days), suggesting either product-stage stability or resource constraints. Pain-point data shows deployment efficiency and cost management are core friction points.
Notable leadership hires: Video Director
与之科技, headquartered in Beijing, develops deep learning deployment and optimization tooling. The company's work spans automated model pruning and quantization, large-scale training technology deployment, and deep learning platform optimization—all targeting faster, cheaper model inference across edge and cloud environments. The technical foundation (PyTorch, TensorRT, C++, ARM, GPU) aligns with a developer-focused, infrastructure-first go-to-market. Current hiring is minimal and heavily weighted toward engineering internships and mid-level roles, reflecting a small, engineering-driven organization.
PyTorch, Python, C++, TensorRT, ARM, DSP, GPU, Linux, plus collaboration tools (Confluence, Jira) and social-media integration (Douyin, Xiaohongshu).
Model deployment optimization (pruning, quantization, TensorRT integration), large-scale training research, platform performance, and internal R&D process improvement. Cost management and deployment efficiency are stated priorities.