Point your AI agent at freehire and let it find you a job. A CLI and an MCP server over the whole job API — no browser.
1. 主导大模型推理性能优化：通过算子优化、模型压缩/量化/稀疏化、推理引擎开发及分布式系统设计等技术手段，全面提升大模型在GPU/PPU等异构硬件上的推理速度、系统效率与部署性价比。 2. 开展软硬协同的深度调优：针对不同架构和规模的AI加速硬件（如GPU、专用AI芯片），对KV Cache管理、内存调度、计算图优化等核心模块进行定制化性能调优，突破算力与访存瓶颈。 3.…
Builds GPS-denied navigation and targeting software for autonomous defense platforms: designing vision navigation and sensor-fusion/state-estimation algorithms (IMU, cameras, LiDAR, radar, GNSS) in modern C++ on Linux, with optional PyTorch/CUDA deployment on embedded Jetson hardware, from simulation through flight testing.
Build and deploy AI models for autonomous transport robots, integrating perception, planning, and edge optimization in real-world industrial settings.
Build and deploy AI models for autonomous transport robots, integrating perception, planning, and edge deployment in real-world industrial settings.
Designs and maintains data pipelines, workflow automation and internal tooling for Sonitec's sensor AI engine, profiling and optimising performance-critical code while running CI/CD, DataOps and MLOps workflows. Day to day involves Python/Java development on Linux/Docker and processing in-field sensor test data.
ML systems engineer at TensorScale who optimizes speed and efficiency across the training and inference stack for video, image, and world-model workloads. Day-to-day work involves writing CUDA/Triton GPU kernels, profiling with Nsight, and building distributed multi-node inference and training systems. Onsite in Menlo Park 5 days a week.
Dyna Robotics is hiring an ML Infrastructure Engineer in Redwood City (in office 5 days/week) to own training infrastructure end to end: scaling distributed training across multi-cloud GPU fleets, building scheduling tooling, and optimizing low-latency inference for robot control. Core stack includes PyTorch, FSDP/ZeRO, Kubernetes, SLURM, CUDA, and TensorRT/Triton.
AI Field Engineer - Enterprise Company: Confidential — generative AI inference and fine-tuning platform Location: US-based, remote-friendly (offices in San Mateo, CA and New York, NY); regular travel to enterprise…
Staff engineer building the ML inference API layer for Cerebras's disaggregated serving system, which pairs GPU-accelerated prefill with ultra-fast decode on Cerebras wafer-scale chips. Day-to-day work spans designing production inference APIs (chat, streaming, tool use, structured outputs), integrating models across backends, and ensuring performance and correctness in Python/Go with C++/Rust sys
A hands-on engineer acting as deep technical partner to strategic AI customers: optimizing LLM inference engines (vLLM, TensorRT-LLM, SGLang), tuning KV cache, parallelism, and quantization to hit performance targets, and guiding fine-tuning/post-training pipelines (LoRA, SFT, DPO, RLHF, GRPO) from POC to production, primarily in Python.
Machine Learning Engineer on Together AI's Inference Engine team, building and optimizing production systems that run large language model inference at scale. Day-to-day work centers on Python and PyTorch, high-performance runtime services, and low-level systems optimization (memory, threading, CUDA/Triton kernels).
Together AI is hiring a Senior ML Engineer to own the model-serving layer of its Voice AI platform in San Francisco — optimizing real-time speech-to-text, text-to-speech, and speech-to-speech inference on GPUs using engines like TensorRT-LLM and SGLang. Day to day: latency/throughput tuning, streaming audio batching, and building model evaluation frameworks.
Staff ML engineer who owns the model-serving layer of Together AI's Voice AI platform — optimizing STT, TTS, and speech-to-speech inference (Whisper, Parakeet, Orpheus, Kokoro) on H100/H200/B200 GPUs with engines like TRT-LLM and SGLang, plus building batching, profiling, and evaluation frameworks. On-site in San Francisco.
Photon is hiring systems engineers to build a next-generation AI inference platform from scratch in Rust — owning batching, scheduling, KV cache management, request routing, and multi-GPU serving. It's an on-site role in Palo Alto for engineers with deep inference-internals knowledge.
1、参与视觉生成/多模态模型（包括文本、图像、视频生成等）在 GPU、ASIC、FPGA 等异构硬件上的推理/后训练加速开发与软硬件结合的性能优化工作，包括但不限于模型量化、attention优化、显存优化、编译优化、计算与通信优化、内存管理以及多卡或多设备的并行推理方案等； 2、在主流深度学习框架（如 PyTorch）基础上，基于GPU/xPU硬件特点，对关键算子进行软硬件结合优化，提升模型运行效率；…
Embedded Engineer in Buntar Aerospace's R&D team integrating onboard electronic systems for UAV platforms — autopilots, onboard computers, navigation, communication links and payloads — and supporting bench and flight testing. Core stack: ArduPilot, MAVLink, embedded Linux (Ubuntu), and interfaces like CAN, UART, I2C, SPI and Ethernet.
Staff-level applied AI engineer at Ivanti who builds data pipelines over Snowflake and unstructured data, ships churn analytics and predictive models, and produces production LLM/agentic AI systems optimized for token cost and latency. Also defines enterprise reference architectures and MLOps/LLMOps standards for other teams.
About the role We are seeking a versatile Full-Stack Algorithm Engineer to join our international product team, focusing on the end-to-end R&D, optimization and production delivery of document intelligence,…
You will build runtime modules that connect compilers, drivers, and ML frameworks for model deployment. You will support PyTorch execution, develop profiling capabilities, extend vLLM, optimize multi-NPU distributed…
We couldn't check your fit for this role — add a CV to your profile to see it next time.
