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. 负责音频生成大模型研发（文生音频、音频编辑、音色克隆、音效合成等），产出高保真、可控的生成方案并推动落地； 2. 研发音频理解大模型，构建跨模态（文本、图像、视频与音频联合）的细粒度音频内容描述体系； 3. 搭建大规模多源通用音频数据集（音乐、语音、环境声等），支撑模型持续迭代； 4. 推动音频大模型在业务场景中落地，解决推理效率、可控性、实时性等工程问题； 5. 跟进行业前沿，保持团队在通用音频大模型领域的技术领先。…
1. 用户偏好采集与建模：设计并构建系统化的用户偏好数据采集方案，深入分析用户行为与反馈信号，建立多维度、可扩展的用户偏好表征与建模体系，为 Reward Model 训练提供高质量数据基础。 2. Reward Model 训练与优化：训练高精度、低偏差的 Reward Model，覆盖创作质量、指令遵循、安全对齐、人类偏好等多个专项维度；持续优化训练策略，有效缓解 Reward Hacking 与分布偏移问题，为下游 RL…
Build predictive world models using generative AI and deep learning to simulate how scenes evolve, training autonomous driving and robotics policies from multimodal sensor data.
Builds and maintains developer-facing tools (APIs, SDKs, dashboards) on top of AI inference/training infrastructure like SGLang and Miles, collaborating with product and research teams to create intuitive interfaces.
Build and optimize production-grade inference services for Apple’s large language, vision, and speech models, ensuring low-latency performance across products like Siri and Photos.
1. 负责自动驾驶模型训练效率与算法迭代效率方向的技术规划和团队管理，持续优化从实验提交、训练执行到评测结果产出的端到端周期； 2. 负责大规模分布式训练性能优化，围绕计算、显存、通信、数据加载、存储等环节定位瓶颈，持续提升训练吞吐、GPU 利用率、扩展效率和训练产出； 3. 建设和优化统一训练能力，包括但不限于分布式训练框架、训练…
Optimize AI training and inference workloads for speed, cost, and efficiency across the full stack, from GPU kernels to distributed systems, using Python, C++, and profiling tools.
Build and deploy agentic AI systems for financial data, focusing on LLM orchestration, retrieval, and scalable workflows to power generative AI applications in finance.
Build and train large language models from scratch for coding assistance, deploying them to production across JetBrains’ developer tools ecosystem.
Build and maintain MLOps infrastructure and pipelines to streamline AI model development, training, and deployment for JetBrains' developer tools and AI-powered IDE features.
1、探索下一代Agent：研究Agent在复杂长程任务（Long-horizon Tasks）中的能力边界；探索Planning、Memory、Tool Use、Multi-Agent、Proactive Agent等方向；构建面向真实生产环境的大规模Agent系统； 2、大模型后训练（Post-training）：基于SFT、Preference Learning、RL（GRPO、PPO等）持续提升模型能力；构建Agent…
ML engineer builds and runs online reinforcement-learning pipelines to improve GigaChat’s post-training, designing experiments, reward signals, and distributed training workflows.
Lead a team building and optimizing scalable ML training platforms on AWS/Kubernetes, focusing on GPU workloads, performance tuning, and Gen AI/LLM pipelines while enforcing enterprise governance and security standards.
团队介绍：Data AML是字节跳动的机器学习中台，为抖音/今日头条/西瓜视频等业务提供推荐/广告/CV/语音/NLP的训练和推理系统。为公司内业务部门提供强大的机器学习算力，并在这些业务的问题上研究一些具有通用性和创新性的算法。同时，也通过火山引擎将一些机器学习/推荐系统的核心能力提供给外部企业客户。 1、扩大推荐模型的行为序列长度、参数规模和训练规模，验证推荐领域的Scaling
Build and improve Avito’s in-house large language models by designing experiments, training pipelines, and optimizing inference for production use.
Build and scale AI/ML infrastructure for autonomous-driving simulations using large foundation models, focusing on distributed training and ML accelerators.
Build and scale AI/ML infrastructure for ultra-realistic multi-agent simulations using large foundation models to train and test Waymo’s autonomous driving systems.
Builds and optimizes large-scale AI infrastructure, including Kubernetes clusters, RDMA networking, and GPU orchestration to improve efficiency and scalability of AI training/inference systems.
Build and deploy large language models for financial workflows, optimizing training and serving to improve agent efficiency and customer experience in banking systems.
Build and optimize large language models for financial services, focusing on LLM-based methods, training pipelines, and production deployment to improve customer workflows and agent efficiency.
We couldn't check your fit for this role — add a CV to your profile to see it next time.