He led AI bootcamps & roundtables – now he’s an Associate Project Manager at Vector.
His advice: “Hit that submit button; you won’t regret it!”
Apply for Applied AI internships by Feb 9: vectorinstitute.ai/programs/int...
He led AI bootcamps & roundtables – now he’s an Associate Project Manager at Vector.
His advice: “Hit that submit button; you won’t regret it!”
Apply for Applied AI internships by Feb 9: vectorinstitute.ai/programs/int...
This isn't a typical internship. You'll contribute to active AI implementation projects and work within Canada's premier AI research ecosystem. 4 specialised roles available.
Apply by Feb 9: vectorinstitute.ai/programs/int...
This isn't a typical internship. You'll contribute to active AI implementation projects and work within Canada's premier AI research ecosystem. 4 specialised roles available.
Apply by Feb 9: vectorinstitute.ai/programs/int...
In this #WiAIRpodcast episode, we speak with @mryskina.bsky.social on neuroscience × AI, evaluation limits, interpretability, and why community shapes better research.
🎬 youtu.be/PQx4IvJR8Bg
In this #WiAIRpodcast episode, we speak with @mryskina.bsky.social on neuroscience × AI, evaluation limits, interpretability, and why community shapes better research.
🎬 youtu.be/PQx4IvJR8Bg
Remember: Your program nominates you. Contact your administrator now about the nomination process
📅 Vector’s deadline: March 18 at 4 PM EST
📋 Info session: February 11 at 12 PM: vectorinstitute.swoogo.com/vsai_2026_pr...
Remember: Your program nominates you. Contact your administrator now about the nomination process
📅 Vector’s deadline: March 18 at 4 PM EST
📋 Info session: February 11 at 12 PM: vectorinstitute.swoogo.com/vsai_2026_pr...
Thank you for making 2025 extraordinary. We'll see you back here on January 5.
Wishing everyone a restful holiday season 🎄
Thank you for making 2025 extraordinary. We'll see you back here on January 5.
Wishing everyone a restful holiday season 🎄
Paper: www.biorxiv.org/content/10.6... #neuroAI #compneuro #neuroscience
@uhn.ca @uhnresearch.ca @vectorinstitute.ai @utoronto.ca @uoftcompsci.bsky.social
Paper: www.biorxiv.org/content/10.6... #neuroAI #compneuro #neuroscience
@uhn.ca @uhnresearch.ca @vectorinstitute.ai @utoronto.ca @uoftcompsci.bsky.social
Steven's story shows how the VSAI opens doors beyond funding.
Ready to write your own success story?
vectorinstitute.ai/programs/sch...
Steven's story shows how the VSAI opens doors beyond funding.
Ready to write your own success story?
vectorinstitute.ai/programs/sch...
Steven connected with Roche at Vector's AI Summit, found their internship posting on Vector Talent Hub, and secured an 8-month role as a Data Science Associate Intern within weeks.
Steven connected with Roche at Vector's AI Summit, found their internship posting on Vector Talent Hub, and secured an 8-month role as a Data Science Associate Intern within weeks.
At Vector, Yisong tackled cultural bias in LLMs with Faculty Member Dr. Vered Shwartz, exploring why LLMs assume North American contexts and how to help them adapt to different cultures
Applications close Jan 12 at 1PM : vectorinstitute.ai/programs/res...
At Vector, Yisong tackled cultural bias in LLMs with Faculty Member Dr. Vered Shwartz, exploring why LLMs assume North American contexts and how to help them adapt to different cultures
Applications close Jan 12 at 1PM : vectorinstitute.ai/programs/res...
From experiment design to DNA engineering, Vector Faculty Members are positioning AI to compress decades of scientific discovery into years.
🎧 The future of research is collaborative. Listen now.
From experiment design to DNA engineering, Vector Faculty Members are positioning AI to compress decades of scientific discovery into years.
🎧 The future of research is collaborative. Listen now.
Vector's contributions at #NeurIPS2025 span the full spectrum of AI research – from ethical data practices to biological reasoning to advanced optimization. 🇨🇦 Congratulations to all our community – we can’t wait to see what’s next!
Vector's contributions at #NeurIPS2025 span the full spectrum of AI research – from ethical data practices to biological reasoning to advanced optimization. 🇨🇦 Congratulations to all our community – we can’t wait to see what’s next!
FreshStack (Jimmy Lin): Holistic framework for building realistic IR & RAG benchmarks on niche domains
Creating the evaluation systems that drive AI forward.
FreshStack (Jimmy Lin): Holistic framework for building realistic IR & RAG benchmarks on niche domains
Creating the evaluation systems that drive AI forward.
STITCH-OPE (Florian Shkurti): Guided diffusion for off-policy evaluation in high-dimensional spaces
CausalPFN (Rahul Krishnan): Pre-trained transformer for zero-tuning causal effect estimation
From correlation to true causation.
STITCH-OPE (Florian Shkurti): Guided diffusion for off-policy evaluation in high-dimensional spaces
CausalPFN (Rahul Krishnan): Pre-trained transformer for zero-tuning causal effect estimation
From correlation to true causation.
Pixel Reasoner (Wenhu Chen): Pixel-space reasoning achieving 84% on V*bench, 74% on TallyQA-Complex
TIRE (Igor Gilitschenski): Subject-driven 3D/4D generation with superior identity preservation
AI that truly sees and reasons visually.
Pixel Reasoner (Wenhu Chen): Pixel-space reasoning achieving 84% on V*bench, 74% on TallyQA-Complex
TIRE (Igor Gilitschenski): Subject-driven 3D/4D generation with superior identity preservation
AI that truly sees and reasons visually.
BioReason (Chris Maddison & Bo Wang): DNA-LLM integration achieving 97% accuracy on disease pathway prediction
ELECTRA (Alán Aspuru-Guzik): Floating orbital approach for efficient molecular charge density prediction
BioReason (Chris Maddison & Bo Wang): DNA-LLM integration achieving 97% accuracy on disease pathway prediction
ELECTRA (Alán Aspuru-Guzik): Floating orbital approach for efficient molecular charge density prediction
The Common Pile (Colin Raffel): 8TB of openly licensed text training a competitive 7B-parameter LLM
What is Your Data Worth to GPT? (Roger Grosse): 6,500x faster influence functions with 5x less memory
The Common Pile (Colin Raffel): 8TB of openly licensed text training a competitive 7B-parameter LLM
What is Your Data Worth to GPT? (Roger Grosse): 6,500x faster influence functions with 5x less memory
Vector's research community wraps up with foundational breakthroughs in data attribution, biological reasoning, and AI optimization. Here's how we're shaping AI's future: 🧵
Vector's research community wraps up with foundational breakthroughs in data attribution, biological reasoning, and AI optimization. Here's how we're shaping AI's future: 🧵
Vector Faculty Members are tackling uncertainty quantification, content safety, deployment monitoring, and robotic reliability – building the foundation for AI we can rely on.
Vector Faculty Members are tackling uncertainty quantification, content safety, deployment monitoring, and robotic reliability – building the foundation for AI we can rely on.
From molecular dynamics to AI safety, our research community is building the scientific foundations for tomorrow's breakthroughs. 🇨🇦
Full list of accepted papers: vectorinstitute.ai/vector-resea...
From molecular dynamics to AI safety, our research community is building the scientific foundations for tomorrow's breakthroughs. 🇨🇦
Full list of accepted papers: vectorinstitute.ai/vector-resea...
OpenCUA (Victor Zhong): Open foundations for computer-use agents with 27K demonstrations
AgentNet-7B achieves 20.1% success on OSWorld, 21.1% on WindowsAgentArena
Building the future of AI agents, openly.
OpenCUA (Victor Zhong): Open foundations for computer-use agents with 27K demonstrations
AgentNet-7B achieves 20.1% success on OSWorld, 21.1% on WindowsAgentArena
Building the future of AI agents, openly.
Ctrl-DNA (Bo Wang): Controllable cell-type-specific regulatory DNA design via constrained RL
RETRO SYNFLOW (Renjie Liao): Discrete flow-matching for accurate retrosynthesis prediction
AI-powered synthetic biology and drug discovery.
Ctrl-DNA (Bo Wang): Controllable cell-type-specific regulatory DNA design via constrained RL
RETRO SYNFLOW (Renjie Liao): Discrete flow-matching for accurate retrosynthesis prediction
AI-powered synthetic biology and drug discovery.
SAFE (Igor Gilitschenski & Florian Shkurti): Multitask failure detection for vision-language-action models
DiffBreak (Yaoliang Yu): Exposing vulnerabilities in diffusion-based purification defenses
Making AI systems safer for deployment.
SAFE (Igor Gilitschenski & Florian Shkurti): Multitask failure detection for vision-language-action models
DiffBreak (Yaoliang Yu): Exposing vulnerabilities in diffusion-based purification defenses
Making AI systems safer for deployment.