Research engineer at the intersection of applied category theory, Bitcoin (UTXO / Script), and coordination systems (protocols, agents).
Themes: closing the semantic gap between off-chain intent and on-chain truth • formal / auditable meta-protocols • UTXO-native design (security through structure, not VM complexity) • relational / geometric semantics for LLMs and for protocol objects (algebraic topology, differential geometry, Galois-theoretic angles) • PI coding agents in the research loop.
Mathematical semantics: I use algebraic topology, differential geometry, and Galois-theoretic perspectives to study semantic structure—both for LLMs (compositionality, limits, and obstructions in high-dimensional learned systems) and for Bitcoin transactions as rigid, auditable state transitions. The parallel is intentional: both settings benefit from invariant-centric explanations and explicit structure, with Bitcoin as the anchored reference case and LLMs as the open, high-dimensional one.
BTSL — Bitcoin Transaction Schema Language — declarative “policy-as-code” layer on top of BIP174/BIP370 (PSBT): schemas describe expected structure, calc + ASSERT encode economic invariants, external Checkers get an explicit predicate set for zero-trust replay.
Release (normative baseline): v1.0.0 — Reference Specification [FINAL] — spec §9.3.1 checker predicates (S-1…A-5), full BTSL_ERR_00…13, implementation guide + checker annex.
Live explorer: BTSL Playground (Maker + Validator paths; implementation is tracked to the FINAL spec over time).
| Focus | Link |
|---|---|
| Formal semantics for Bitcoin metaprotocols (Ordinals / Inscriptions, semantic-gap framing) | formal-metaprotocol-semantics |
| Curated Bitcoin research notes (socio-technical layer, permanence, censorship resistance) | bitcoin_research |
| Technical posts & French translations of Bitcoin research (ecosystem outreach) |
Experimentation with Model Context Protocol (MCP)-driven tool use, multi-agent setups, and “sovereign” agent design—research questions: coordination surfaces, trust boundaries, and anchoring assertions in UTXO-time rather than mutable off-chain state. Exploratory deployments across social surfaces and Bitcoin-adjacent tooling (e.g. lasereyes).
- Education: Computer Science (M1) — AMU & ULB • Mathematics & Applications (Master, research & teaching track) — AMU
- Industry: CTO (Tokenomeme, infra / GCP) • Blockchain dev — Juneo (Avalanche-class fork; e.g. juneojs-examples)
- Community: Speaker supervisor — UrLab (ULB hackerspace) • CIVIS ambassador (AMU) • steward / contributor — DAO Brussels (manifesto)
Formal methods
Core dev
Bitcoin / PSBT
Workflow
Bitcoin’s consensus layer is a long-horizon anchor for truth and coordination. I work on explicit semantics and verifiable artifacts (schemas, proofs, static pipelines) so that things built above that anchor stay honest when humans and agents automate the middle layers—without smuggling account-model assumptions back into UTXO-native designs.
Open to collaboration and critique on Bitcoin formalization, PSBT / BTSL, category-theoretic protocol semantics, mathematical semantics of LLMs, and agent coordination.