Video by FINOS via YouTube
Andres Rojas (Director of Applied AI Projects at the Vector Institute) explores why the true "moat" in financial services is no longer the AI algorithm itself, but the shared foundations and high-quality data used to train it. He provides a roadmap for staying at the state-of-the-art in a world where AI research is accelerating by 30% year-over-year.
🇬🇧 Join us in London! Catch the latest on Applied AI at OSFF London on June 25, 2026: https://hubs.ly/Q041YV9Z0 (Use Code: 26YTOSFFLN20C)
🕒 Timestamps:
0:00 Setting the Context: Research vs. Application
0:50 The Key Message: Shared Foundations in AI
1:15 Why Algorithms Aren’t IP: The Training Layer
1:45 Transformers and Model Weights
3:05 The Speed of Change: 100,000 Papers a Year
4:45 Proof: Collaboration Results with 200 Companies
5:55 Trust Through Community Validation
6:45 Closing Takeaways: Invest in Your Edge
📊 The Problem: The Myth of the Proprietary Model Many organizations believe their competitive advantage lies in protecting their specific machine learning algorithms as Intellectual Property (IP). However, in modern AI, the logic (like Transformers) is inherently open-source and shared globally; trying to keep these foundations secret only slows down exploration and increases the cost of staying relevant.
🏗️ The Solution: Focusing on the "Differentiating" Layers
Andres Rojas argues that the value shift in AI requires a new strategic focus:
* The Training Moat: IP is created only when open algorithms are trained on your specific, high-quality datasets to create calibrated weights.
* Accelerated Exploration: Collaborating at the foundational layer allows firms to digest the "firehose" of 2,000+ AI papers published monthly and move to production faster.
* Assurance Over Blind Trust: Open foundations allow for community-led validation, building the "Three Lines of Defense" for AI that proprietary "black boxes" cannot provide.
⚙️ Why This Matters for Financial Engineering
* Efficiency Through Sharing: Collective co-development has saved partners over 250,000 hours in development time by mutualizing infrastructure.
* Lead on Openness: Just as AI was launched in Canada, the local financial community can lead by building on verified, open foundations rather than general algorithmic secrets.
The takeaway: Invest your edge in data, domain expertise, and governance—not the algorithm. Andres Rojas proves that the fastest path to trustworthy AI is through shared foundations.
🌐 More about FINOS: https://www.finos.org/
📧 Join our newsletter: https://www.finos.org/sign-up
🎙️ Listen to our Open Source in Finance Podcast: https://www.youtube.com/@FINOS/podcasts
LinkedIn: https://www.linkedin.com/company/finosfoundation
#FINOS #OSFFToronto #VectorInstitute #AIStrategy #DeepLearning #OpenSource #FinTech #AppliedAI