02 / RESEARCH TO VENTURE
From frontier research to global products.
Following the people who connect research depth, engineering ability and product judgment — across the next generation of AI-native companies.
INDEPENDENT SELECTION
SIGNAL List / Talent & Research
Selected research teams, technical disciplines and university ecosystems across the path from AI research to products.
Model & research teams
OpenAI
Post-training, reinforcement learning, agents and inference.
Anthropic
Model research, reinforcement learning and ML systems.
Google DeepMind & Google Research
Multimodal AI, robotics, model efficiency and research engineering.
Meta
AI research, model systems and recommendation infrastructure.
Microsoft Research Asia
Multimodal research, intelligent interaction and applied AI.
ByteDance Seed
Foundation models, multimodal AI, world models and AI infrastructure.
Qwen & Alibaba model teams
Foundation models, post-training and model infrastructure.
NVIDIA Research / GEAR
Embodied AI, robot learning, simulation and synthetic data.
Tencent Hunyuan
Language, image, video and 3D model research.
DeepSeek
Foundation-model research and open-source AI.
MiniMax
Language, video and speech models with AI-native products.
Technical roles & disciplines
Post-training & RL
Reasoning, reward modelling and model adaptation.
Agent research & evaluation
Tool use, task environments and human–AI interaction.
ML systems & inference
Serving, kernels, compilers and distributed training.
Data & retrieval systems
Data curation, retrieval and synthetic data.
Multimodal & generative media
Vision, video, audio, 3D and world models.
Robotics & edge AI
Robot learning, simulation and deployment on real devices.
Applied ML & product engineering
Applied science, ranking and computational photography.
PhDs, postdocs & open-source builders
Independent research, working prototypes and sustained contributions.
Universities & research ecosystems
Stanford
Hazy Research, CRFM and Physical and Spatial AI.
UC Berkeley
Sky Computing, BAIR, RAIL and the vLLM community.
Carnegie Mellon University
Robotics Institute and machine learning research.
MIT
CSAIL, robotics and efficient model research.
ETH Zürich
Robotic Systems Lab and robotics spin-offs.
EPFL
Robotics, control and intelligent hardware.
Oxford
OATML: machine learning, reliability and uncertainty.
UCL
Gatsby: machine learning and computational neuroscience.
Mila
AI research and a research-to-venture ecosystem.
Tsinghua University
AIR and embodied intelligence / robotics research.
Peking University
Foundation models, multimodal AI and embodied intelligence.
Shanghai Jiao Tong University
AI, computer vision and robot learning.
Zhejiang University
Robotics research and engineering translation.
HKUST / XbotPark ecosystem
Robotics, hardware and engineering entrepreneurship.
National University of Singapore
AI Institute and NUS AI Lab.
Nanyang Technological University
S-Lab: vision, language, reinforcement learning and systems.
An independently curated list of institutions, teams and research disciplines. Inclusion does not indicate a SIGNAL partnership.
01 / PERSPECTIVE
Research depth. Product judgment.
AI brings research, engineering and product development into closer conversation. A promising idea becomes more meaningful as a team learns to make it reliable, useful and accessible to others.
SIGNAL follows the people making that transition: researchers, engineers and builders who develop their own questions, build tangible work and keep learning from the world outside the lab.
02 / WHAT WE LOOK FOR
What we pay attention to.
01 / ORIGINALITY
Independent questions
The ability to choose meaningful problems, develop a clear technical point of view and contribute work that others can build on.
02 / ENGINEERING
Engineering ownership
The ability to carry an idea into a working system, understand its constraints and improve it through repeated delivery.
03 / PRODUCT
An understanding of users
Curiosity about who uses the technology, what changes in their work or lives, and how feedback can shape the next iteration.
04 / GLOBAL
Global ambition
The openness to work across communities, learn from different markets and build for people beyond a familiar environment.
03 / OUR FOCUS
Three frontiers we stay close to.
Across university research, technology teams and open-source communities, we follow how ideas develop into systems and products. Our attention centres on AI Agents, AI Hardware and the data infrastructure that supports them.
AI Agents
Reliable action, memory and context — and the feedback that turns model capability into useful workflows.
AI Hardware
The connection between learning systems, physical engineering and the experience of everyday use.
Data Infrastructure
The foundations for persistent context, real-time knowledge and dependable AI-native systems.
Stay close to the person behind the work.
Potential develops over time. Through ongoing conversations, shared exploration and honest feedback, SIGNAL builds the context to understand new directions as people and their ambitions evolve.
SIGNAL PERSPECTIVES