Top Tech Jobs & Startup Jobs

Lead the development of AI-powered employee productivity experiences across web and mobile. Architect scalable services and UIs, integrate LLM and vector-search capabilities, connect HR and productivity platforms, and establish cloud-native engineering practices. Own system design, code reviews, observability, security, accessibility, and technical documentation while mentoring engineers and guiding cross-functional teams toward reliable global products.

Design and optimize agentic AI systems for NVIDIA’s CUDA ecosystem, including agent architectures, multi-agent workflows, tool use, memory, orchestration, training, verification, reward, and evaluation systems. Co-design AI systems across models, inference, compilers, runtimes, libraries, kernels, and GPUs. The role requires strong C/C++, Rust, and Python skills, GPU programming expertise, and experience with agent systems, reinforcement learning, or modern AI inference.

Design and deploy test methodologies for advanced NVIDIA silicon, including DFT architectures, ATPG verification, coverage analysis, pattern generation, and ATE bring-up. Partner with EDA vendors, CAD, VLSI, and product engineering teams to develop scalable test solutions. Lead complex scan-test lifecycles, support advanced chiplet and 3D designs, and mentor junior engineers.

Lead developer-focused performance optimization for RTX and DGX SoC platforms: collaborate with partners and internal teams to optimize CPU, AI, and 3D-graphics workloads, profile and debug CPU/GPU performance, influence next-gen SoC design, and mentor junior engineers. Occasional travel for conferences and partner site visits.

Develop Linux kernel and user-space power and thermal management software for NVIDIA Tegra and embedded platforms. Optimize performance, efficiency, latency, compute and memory utilization for edge AI and robotics workloads. Collaborate with hardware, software, and power architecture teams on platform estimation and optimization. Requires strong C, Linux, embedded systems, ARM, and system-level debugging experience.

Supports and root-causes complex customer issues involving NVIDIA networking and AI cluster technologies. Develops bug fixes, workarounds, diagnostic tools, and performance-tuning recommendations in collaboration with R&D and support teams. Works with InfiniBand, NVLink, Spectrum-X, and high-performance lossless networks, analyzing logs, traces, metrics, and source code. Provides customer support and may work occasional weekends or holidays.

Design, implement, and operate large-scale observability and telemetry platforms focused on performance, reliability, monitoring, logging, and alerting. Manage services throughout their lifecycle, including system design, deployment, capacity planning, automation, incident response, and postmortems. Build infrastructure tools and frameworks for private and public cloud systems, improve availability and latency, and participate in on-call support for production environments.

Leads the development and production deployment of deep learning perception systems for autonomous vehicles. Manages engineering managers, researchers, and software engineers; directs architectures for multi-sensor fusion and real-time 3D world modeling; oversees the machine learning lifecycle from data generation through NVIDIA DRIVE deployment; and partners across planning, mapping, hardware, and safety teams while ensuring automotive quality standards.

Develop and optimize GPU-accelerated deep learning inference software for LLMs, multimodal, and generative AI models. Contribute features to vLLM, SGLang, FlashInfer, and NVIDIA inference libraries; implement algorithms, tune performance across GPU architectures, and improve large-scale model serving. The role requires strong C/C++ software engineering, collaboration across framework teams, and expertise in profiling, debugging, performance modeling, GPU programming, or multi-GPU communication.

Contribute to power estimation models for GPUs, explore architecture for energy efficiency, analyze performance versus power, and employ machine learning for modeling power and performance metrics.