Alec James · Head of Product, Nexus

How do you design systems that can't be corrupted?

Six years across climate systems, AI infrastructure, and global markets. Not a career path — a single compounding thesis about the hardest problem in design.

Recognized · Published · Backed

The New YorkerFeatured
United NationsAcademic Impact Fellow
Duke UniversityNicholas School Grant
Lightspeed FactionInvestor in companies I've built
Pantera CapitalInvestor in companies I've built
DragonflyInvestor in companies I've built
01The first question

1 of 3 selected from 700+ global entries

UN Academic Impact Fellowship

2019 – 2022

What if software could save the planet?

I was studying at Duke when climate change clarified itself as the defining coordination failure of our era. Not a science problem — a systems design problem. I built a climate action webapp to drive high-impact behavioral change, and it got noticed: selected from over 700 entries for the UN Academic Impact Fellowship, backed by grants from Paul English (Kayak) and Duke's Nicholas School of the Environment. Then Bill McKibben featured it in The New Yorker. First real proof that well-designed software moves people at scale.

02A new primitive

United Nations Development Programme

Presented applications to global impact

2021 – 2023

Trust without corruption.

A new design primitive emerged that changed my mental model completely: programmable trust. Not just a new financial instrument — a new infrastructure layer where fairness, composability, and verifiability become engineering choices rather than institutional promises. I built on it. Worked on cultural heritage preservation across tourism-dependent economies. Presented its applications to impact at the UNDP. Each project deepened the same conviction: the most important lever for human coordination is architecting systems where defection is structurally costly.

03The alignment problem

First AI Settlement Oracle

Original research — neutral AI for dispute resolution

2024

What if AI could be a more objective arbiter of truth than humans?

As AI capabilities accelerated, the question of who controls the most powerful models became urgent. I joined ORA Protocol to pursue a specific thesis: distributed AI infrastructure as a lever for alignment, cost reduction, and access. Published original research on AI Settlement Oracles — a new architecture where neutral AI replaces human consensus in dispute resolution, removing economic incentives from the truth-finding process. The question underneath it matters far beyond prediction markets.

Read the research paper
04Community-owned AI

World's first community-owned AI model

Decentralized · Tokenized · Live

2024

The world's first decentralized, tokenized foundational model.

I led the launch of the world's first community-owned, tokenized foundational AI model. The thesis: AI is only aligned if its ownership is aligned. Concentrating the most powerful models inside a handful of labs is a governance failure waiting to compound. Community ownership changes the incentive structure at the root — ensuring equity, diversity of opinion, and shared stake in outcomes baked into the model itself, not bolted on later as policy.

05The current bet

Nexus: live

Head of Product

2024 – present

Every asset. For everyone. No gatekeepers.

The financial system is the last major infrastructure that still runs on permission. US stocks, FX currencies, private markets, credit — access to these assets is determined by where you were born, what bank you have, and who your broker is. Nexus removes those constraints. A financial layer that is permissionless, programmable, and flexible enough to meet any market participant wherever they are. I'm Head of Product, working across the full stack to make market and asset accessibility a universal right.

Lightspeed FactionPantera CapitalDragonfly

Advisory

Frontier companies I've helped build.

I've acted as a part-time advisor and contributor for companies across AI and frontier technology since 2022 — working across product strategy, go-to-market, and technical direction.

Shuffle Infrastructure

AI Infrastructure
2025 – Present

AI hyperscaler infrastructure. Building the compute layer for the next generation of AI workloads.

AI Seer

AI · Media
2025 – 2026

The most accurate real-time fact-checking AI system for multimedia. Winner, TIME 200 Best Inventions 2024. Exploring decentralized incentive mechanisms to accelerate fact-checking and model feedback at scale.

Tim DraperDraper AssociatesAmand Ventures

Lvmodel

AI · Infrastructure
2024 – 2025

LLMs and GNNs optimized for high-throughput network environments. Consumer applications built on these foundational models: smarter block explorers, deep crypto research tools, and more.

Helios Labs

Fintech · Climate
2022

Democratizing global solar project investment using stablecoins. Making infrastructure-grade green energy accessible to retail investors worldwide.

What I believe

Corruption is always a design flaw, not a moral failure.

The best coordination mechanisms make defection structurally impossible.

AI without diverse ownership is just a faster path to the same power concentrations.

Every system that gatekeeps access is a system waiting to be disrupted.

The most important products make the right behavior the easy behavior.

The frontier is wherever the hardest problems haven't been solved yet. That's where I want to be.

For investors & builders

Working on hard problems at the frontier of AI, financial infrastructure, or market access?

I've been at this intersection for six years. Happy to talk thesis, co-build, or advise.

alecjames12@gmail.com ↗

For recruiters & operators

Building a product team at the frontier of technology? Industry doesn't limit me.

I lead from first principles, ship fast, and kill features ruthlessly. Let's talk.

@alecwaj on X ↗