Charlie Thompson

charlie_thompson@protonmail.com · San Francisco, CA · charliethompson.lol · contextengine.sh

Summary

Researcher and engineer building open, public-interest infrastructure for collective decision-making. Creator of Context Engine, an open-source (MPL-2.0) platform for AI-assisted deliberation in large groups. Longtime RadicalxChange community contributor on plural and quadratic governance; Cornell CS (cryptography, NLP, AI).

Focus Areas

• AI-assisted deliberation at scale: finding consensus while preserving viewpoint pluralism and preference intensity.
• Representation and fidelity: evaluating whether AI agents and AI syntheses faithfully represent the people they speak for.
• Public-interest cryptography: digital voting, private proofs, proof-of-humanity, and verifiable agreements.

Experience

• Created Context Engine, an open-source (MPL-2.0) platform for structured deliberation in large groups — public/private surveys and discussions, response clustering, and AI-assisted synthesis for humans and AI agents; developed full-time since 2023 and open-sourced in 2026 (3,500+ commits, 58 tagged releases).
• Deployed at community deliberation events — including structured watch parties for the 2025 Weyl–Yarvin debate, at hundreds to low thousands of concurrent participants — with prototypes tested in the RadicalxChange community.
• Ran “Agent Village Wrapped” at Edge Esmeralda 2026, an early public evaluation of how faithfully personal AI agents represent their principals, scoring fidelity and calibration against participants’ own answers.
• Curated and published an open AI-discourse corpus — question sets spanning AI safety, governance, and policy — with tooling for comparing human and model positions.
• Built for non-technical participants: voice input, plain-language flows, and passkey onboarding that hides key management; cryptography for privacy and eligibility — selective encryption, participant-controlled access to sensitive responses, and non-transferable group credentials.
• Selected for Protocol Labs’ d/acc Residency at Edge City Patagonia (2025); on the proposed program for a demo session at EDDY 2026 (European Digital Democracy conference, Vienna); presented “Social Infrastructure for AI on Ethereum” at Zuzalu AI×Crypto (2023).

• Identified and responsibly disclosed a cryptographic protocol vulnerability; recognized in the QRL Bug Bounty Hall of Fame.

• Contributed to an Ethereum vault prototype using XMSS, a stateful hash-based post-quantum signature scheme.

• Worked on a Jenkins CI pipeline; received a full-time return offer.

• Research support on “A Flexible Design for Funding Public Goods” (Buterin, Hitzig & Weyl), the quadratic-funding mechanism.
• Helped organize the inaugural RadicalxChange conference and connected Democracy Earth with RadicalxChange ahead of the Colorado State House quadratic-voting pilot.

• Designed an early proof-of-human smart contract using natural-language (Winograd Schema) challenges resistant to AI automation.
• Published “Quantum Computers Pose a Credible Threat to the Security of Bitcoin”, an analysis of quantum risk to Bitcoin and ECDSA.

Skills

Software: TypeScript/JavaScript, React, Node.js, Cloudflare Workers, Solidity, Foundry · AI: Anthropic, OpenAI, and OpenRouter APIs; multi-provider orchestration; AI-assisted analysis · Cryptography: passkeys, selective encryption, soulbound credentials, Arweave, XMSS

Education

Cornell University — B.A., Computer Science, 2021 · Focus: Cryptography, Natural Language Processing, AI