← Back to roles

微信-语音全双工对话大模型预训练算法工程师

Tencent Technology (Shenzhen) Co., Ltd.

北京Posted Aug 18, 2026
Apply on Tencent Technology (Shenzhen) Co., Ltd.

Overview

Open roles
2,267
At this company
See all
Added in 30 days
501
At this company, still open

Skills named in this posting

  • PyTorch / TensorFlow
1.负责语音全双工对话大模型的架构设计与预训练,构建边听边说、实时打断、自然接话和连续对话能力; 2.负责音频理解模型预训练,提升语音内容、说话人、情感语气、副语言信息、环境声音及多人重叠语音的理解能力; 3.负责音频生成模型预训练,研究基于 Audio Token、Neural Codec、自回归、Diffusion 或 Flow Matching 的语音生成方案,提升音质、自然度、情感表现和生成稳定性; 4.负责全双工交互行为建模,包括回复时机、用户打断、抢话与让话、Backchannel、重叠语音和流式生成; 5.负责大规模音频预训练数据建设,包括数据清洗、去重、质量过滤、伪标注、Token 化及数据配比; 6.负责模型规模化训练、效果分析及训练稳定性优化,推动预训练模型的后训练、评测和产品落地。 1.计算机、人工智能、语音信号处理等相关专业,硕士及以上学历; 2.具备全双工语音对话模型预训练经验,参与过端到端 Speech-to-Speech、Audio LLM、Omni Model 或类似模型研发; 3.具备音频理解模型或音频生成模型预训练经验,能够负责模型架构、数据建设、训练策略和效果分析; 4.熟悉全双工对话核心问题,包括 Turn-taking、End-of-turn Detection、User Interruption、Overlapping Speech、Backchannel 和 Streaming Generation; 5.熟悉 Audio Encoder、Conformer、Transformer、Audio Tokenizer、Neural Codec、自回归、Diffusion 或 Flow Matching 等技术; 6.熟练使用 PyTorch,具备大规模分布式训练经验,熟悉 DeepSpeed、Megatron-LM、FSDP 等训练框架; 7.具备较强的实验设计、问题定位和跨团队协作能力。

Apply on the employer's site

Original posting: careers.tencent.com

Apply on Tencent Technology (Shenzhen) Co., Ltd.

RISC-V ecosystem

Tencent Technology (Shenzhen) Co., Ltd. in the RISC-V ecosystem

RISC-V International: Premier TSC

Tencent Holdings Ltd. is a Chinese multinational technology and entertainment conglomerate and holding company headquartered in Shenzhen.

  • Tencent joined RISC-V International as China pushed for open-instruction-set alternatives (widely reported)
  • Its TencentOS Tiny IoT operating system is used on RISC-V MCUs, e.g. tutorials pairing TencentOS Tiny with the GD32V RISC-V MCU
  • Also participates in Chinese RISC-V community events such as the RISC-V Forum in Shenzhen

Same employer

View all 2,267