Valkyrie AI & Investment Frontiers Series: Session 1
⭐ About Valkyrie
Join us for an in-depth conversation with top allocators, foundational builders, researchers, and VCs focused on frontier technology and venture-scale innovation. Valkyrie is a premier venture capital firm investing across three high-impact pillars:
AI Infrastructure — scaling next-generation intelligence, compute architectures, and advanced agentic frameworks.
Energy — addressing the power and sustainability bottlenecks behind modern AI workloads.
Crypto — backing decentralized protocols, data sovereignty platforms, and secure cryptographic rails. As a high-conviction partner, Valkyrie bridges elite research with venture-scale execution — supporting founders across fundraising, customer acquisition, strategic partnerships, and global go-to-market.
Join us for more details: https://www.valkyriefund.xyz/
Click here for more slides and recordings of previous sessions: https://decil.org/syllabus/
🎙 Panel 1: The Next Frontier of AI Systems
Compute · Model Provenance · Efficient Inference · Evaluation · Human-Centered AI What comes after prompt engineering? As AI moves beyond single-model applications, the next bottleneck is becoming infrastructure: compute efficiency, inference cost, model routing, evaluation, provenance, deployment, and continuous improvement. This panel brings together researchers and builders working across open-weight model development, distributed AI systems, efficient inference, human-centered scientific AI, and embodied intelligence. The conversation will cover:
How open-weight models evolve through fine-tuning, merging, distillation, quantization, and re-release.
Why model provenance and lineage matter for trust, compliance, reproducibility, AI safety, and future model marketplaces.
How compute, memory bandwidth, cloud-edge routing, and distributed systems shape the economics of AI deployment.
When multi-agent systems outperform single-model or router-based approaches.
How human-centered AI can support high-stakes domains such as environmental sensing, flood prediction, and disaster response.
What students, executives, founders, and VCs should pay attention to in the next phase of AI.
Panel research collaboration: Model OS / Model Development Initiative
Initiative link: https://www.cs.cmu.edu/~tianqinl/modeldev.html
Moderator
Fernando Jia — Builder of Model OS; Lead, Model Development Initiative
Focus: Multi-model infrastructure, model routing, evaluation, model provenance, and AI systems.
Speakers
Xin Luo — Research Fellow, Model Development Initiative; ML Scientist, Expedia
Focus: Open-weight model provenance, model lineage, production ML, and model evaluation.
Affiliation badges: Expedia Group · Model Development Initiative
Jieyu Eric Lin — Research Scientist, Meta
Focus: Efficient foundation models, distributed ML systems, edge inference, multi-agent learning, and AI infrastructure.
Affiliation badge: Meta
Yitong Li — Stanford HAI Graduate Fellow; Research Fellow, Model Development Initiative
Focus: Human-centered environmental AI, interpretable scientific modeling, flood prediction, and equity-aware AI systems.
Affiliation badges: Stanford HAI · Stanford University · Model Development Initiative · NUS
🌐 Model Development Initiative
The Model Development Initiative is a research and builder network focused on the next generation of model systems. The initiative brings together researchers and builders working on model development, model provenance, evaluation, efficient inference, trustworthy AI systems, open model ecosystems, and embodied intelligence. Our broader network includes researchers and builders connected to CMU, Stanford HAI, Berkeley BAIR/RDI, MIT, Meta, OpenAI, Google, Microsoft, and other leading AI ecosystems. Students, researchers, founders, executives, and investors interested in AI infrastructure, model provenance, evaluation, efficient inference, trustworthy AI systems, and embodied intelligence are welcome to connect with us after the session.
Learn more: https://www.cs.cmu.edu/~tianqinl/modeldev.html
🎙 Founder Keynote
Cosimo Spera, Founder & CEO of MinervaCQ, will give a brief keynote on Certified Resolution: Proof-Carrying AI. The keynote will explore why enterprise agentic systems require authorization, governance, verifiable execution, and certifiable answers — especially as AI agents move from demos into regulated, high-stakes workflows.
🎙 Panel 2: Building AI-Native Companies
(Representing the cutting edge of Valkyrie’s portfolio) How are founders turning AI capabilities into products, companies, and lasting competitive advantages? Founders from AI-native startups will discuss what they are building, the problems they are solving, and what they have learned while taking AI products from concept to market. The conversation will cover: product development, technical differentiation, proprietary data, customer adoption, go-to-market strategy, fundraising, and early-stage execution. Featured speakers and startups:
Cosimo Spera, Founder & CEO of MinervaCQ — https://www.minervacq.com/
Mengfan Zhang, Founder & CEO of HealMint — https://healmint.org/
Yunan Ma, Operations Lead of Zillion Network — https://zillion.network/
Jeff Yen, CEO of Etherdyne Technologies / ETI — https://www.etherdyne.net
📅 Date & Location
Date: Saturday, July 4, 2026
Time: 2:00–3:30 PM
Location: 2933 Bunker Hill Lane, Suite 205, Santa Clara, CA 95054 Additional speaker updates and research collaborators will be shared on this page.