RAYNEO LARGE MODEL ALGORITHM DEPT

Building Future Spatial Agents

We are the Rayneo Large Model Algorithm Department.
Responsible for building Cloud-Edge Synergy AI Platforms for smart glasses, and constructing a series of Intelligent Agents.
Empowering glasses with true cognitive abilities through voice interaction and multimodal understanding.

Join Us
Meet the Team
# On-Device ASR+NLU # Intent Recognition # Proactive Memory # Spatial Intelligence

Team Leader

Ada Cheng

Sijie (Ada) Cheng

Ph.D. Candidate @ Tsinghua Univ.

Hiring

Looking for full-time employees and interns for AI in smart glasses. Open for research cooperation.

GitHub Email
ID: AdaChengSijie

Research Focus

I am a second-year Ph.D. candidate at Tsinghua University, advised by Prof. Yang Liu. Previously received Master's from Fudan University.

"Egocentric Foundation Model represents the next generation of Foundation Models."

Egocentric Understanding
EgoThink | VidEgoThink
On-Device Training
OpenChat | ConvLLaVA
Implicit Knowledge
FedGEMs | StableToolBench

Experience

Rayneo

Intern (Jan. 2025 - Present)

Manager: Hongwei Li (CEO)

Tencent (Robotics X)

Research Intern (2024)

Mentor: Lei Han

01.AI Company

Pre-training Group Intern (2023-2024)

Mentor: Xiangang Li

Sinovation Ventures

Investment Intern (2023)

Shanghai AI Lab

NLP Group (2022)

Selected Awards

  • CAST Young Talents Project 2025 Ph.D. Program - China Computer Federation
  • Outstanding Paper Award MFM-EAI Workshop @ ICML 2024
  • National Scholarship China, 2021-2022
  • Outstanding Master’s Thesis Award (Shanghai Computer Society)
  • 1st Place, Tencent Basketball Association (2024)

Recent Talks

  • 2024.11 - Core Competitiveness of Scientific Research in the Era of Large Models
  • 2024.09 - EgoThink @ ZhiDX
  • 2024.03 - OpenChat @ Max-likelihood Community

Team DNA

  • ▹ Young & Geeky: The youngest team in the company, with high talent density and a geeky atmosphere.
  • ▹ Cloud-Edge Synergy: Responsible for building AI platforms for smart glasses and constructing intelligent agents.

Our Goals

Fast

Fast Efficiency

Complete the "Cognition-Implementation-Innovation" loop rapidly.

Adapt

Strong Adaptability

React instantly to technological breakthroughs and adjust course.

Good

Great Culture

Mutual achievement, AI for Good, changing the world.

Team Gallery

Team Building
Holiday Spirit
Team Dinner
Team Dinner
Office Life
Office Life

/ Research Directions

Overview

Our research aims to bridge the gap between the physical world and digital intelligence. By building Egocentric Multimodal Large Models, we empower smart glasses to "understand" the world; through Efficient On-Device Training & Inference, we ensure intelligence is accessible and privacy-secure; by exploring Implicit Knowledge, we enable AI to possess intuition and common sense similar to humans.

Egocentric Understanding

  • ● Achievement 1: Efficient spatiotemporal modeling for egocentric video.
  • ● Achievement 2: Intent recognition for dynamic object interaction in complex scenes.
  • ● Achievement 3: EgoThink: Evaluating the first-person thinking capability of multimodal large models.

On-Device Training

  • ● Achievement 1: On-device LLM training architecture under extremely low power consumption.
  • ● Achievement 2: Model quantization and compression techniques tailored for smart glasses.
  • ● Achievement 3: Cloud-Edge collaborative continuous learning mechanisms.

Implicit Knowledge

  • ● Achievement 1: Mechanism analysis and suppression methods for large model hallucinations.
  • ● Achievement 2: Extracting commonsense knowledge graphs from unstructured data.
  • ● Achievement 3: Research on rapid inference paths simulating human intuition.
We Are Hiring

Join Us, Create the Future

We are looking for partners passionate about AI, Computer Vision, and AR.
We offer competitive compensation, massive GPU computing support, and the opportunity to work with top talent.

Social Recruitment

Full-time Positions

AI R&D Engineer

Description

Responsible for large model capability building, AI algorithm implementation, backend system architecture design, and supporting high-concurrency AI services.

Requirements

Master's degree in CS or above; familiar with microservices architecture (MySQL/Redis/Kafka/Docker); proficient in Python/Java/Golang; experience in cloud development and RAG/Agent implementation.

Intelligent Speech Algorithm Engineer

Description

Optimize core ASR technologies, model iteration, solve wake-up and noise reduction issues; optimize end-to-end voice interaction on devices.

Requirements

Master's in CS/EE or above; proficient in C++/Python; deep understanding of ASR (Conformer/Whisper) principles; familiar with PyTorch/TensorFlow.

Audio Algorithm Engineer (Frontend)

Description

Develop audio AEC algorithms, meet echo cancellation metrics; design Beamforming algorithms; optimize VAD.

Requirements

Proficient in signal processing; familiar with beamforming; proficient in Matlab/Python and C/C++ porting; experience in embedded/DSP optimization.

Speech System Engineer

Description

Responsible for AR/AI glasses audio link design; lead audio risk calibration; coordinate DSP/MCU resource allocation.

Requirements

Experience with consumer audio products; proficient in AEC/Beamforming/VAD algorithm implementation; familiar with Qualcomm/Hisilicon platforms.

Large Model Algorithm Engineer (Memory)

Description

Develop scenario-based memory algorithms, improve personalized storage and retrieval; optimize user profiling and interaction experience; build memory data pipelines.

Requirements

Master's in CS/EE or above; proficient in Python/PyTorch; experience in LLM fine-tuning, RAG, and Agent-memory projects.

Chat Now
Campus Recruitment

Intern Positions

Large Model Algorithm Intern

Description

1. RAG: Build/optimize RAG systems, vector retrieval, KB construction.
2. Intent Recognition/NLP: Build intent classification models, optimize Prompts.
3. AIGC/Multimodal: Participate in text-to-image, multimodal generation model training.

Requirements

Master's/Ph.D. student; proficient in Python; familiar with Transformers/LLM principles; experience in RAG/NLP/Agent; top conference papers (ACL/CVPR etc.) preferred.

Speech Algorithm Intern

Description

Continuously optimize ASR and TTS systems; track cutting-edge technologies; responsible for cloud service architecture and on-device algorithm optimization; develop voiceprint recognition.

Requirements

Bachelor's in CS/EE or above; proficient in C++/Python; deep understanding of ASR/TTS principles (Conformer/Whisper/FastSpeech); familiar with TensorFlow/PyTorch.

AI Efficiency Intern

Description

Develop AI Agent applications; design AI Coding implementation plans for R&D; screen AI efficiency tools; support tool usage.

Requirements

Bachelor's or above; understand LangGraph/AgentScope; familiar with AI Coding tools like Cursor/Claude; strong interest in AI tools.

Chat Now