Thesis
Secrecy between peers, visibility for those with a legal mandate. Not by breaking in — by design.
Privacy that doesn't blind the regulator.
I work on the problem of getting institutions that don't trust one another to agree on the same record, without exposing themselves to each other and without leaving the supervisor in the dark.
I built Bywise — a blockchain written from scratch, not a fork, with a data-sharding architecture that gains capacity with every new node instead of losing it. The code is open source under the MIT licence.
Out of it came ChainXS, developed in a joint venture with Bloxs for the Brazilian financial market. There I designed an auditable privacy layer for a Regulated Liability Network: each user receives fresh addresses for every operation, peers cannot link them to one another, and whoever holds a legal mandate sees the whole tree without being able to sign anything on anyone's behalf. The key to the statement, not the key to the vault.
Before that came fifteen years on the same problem in other clothes — from restaurant tills that had to agree in real time to traceability in logistics, precious metals and exchange proof-of-funds.
Provenance
Every number on this page arrives with a label. Paper accepts anything; engineering does not.
-
Measured 2024
1,000 TPS
Demonstrated record for ChainXS, with declared hardware, in a paper co-authored with Professor Henrique Gomes de Moura (UnB). The proof of concept ran at 100 TPS on a laptop with a 6th-gen i7 and 8 GB of RAM.
-
Theoretical 2021
17 million TPS
Upper bound of the slice architecture — correct in its premises and innocent about the real world: it models neither cross-continent latency, nor disk that can't keep up, nor a home node on a bad connection. Never demonstrated.
-
Marketing 2018
1,000,000 TPS
A promise posted on a cryptocurrency forum, with no whitepaper and not a single test. The first reply I got was a skeptic asking where the document was. He was right.
I keep the three lines up on purpose. The distance between the bottom one and the top one is what I learned, and it's the only credential that survives an audit.
Articles
Privacy requirements in regulated infrastructure, written by someone who took the beating from them.
-
Lien reconciliation: the privacy requirement that survives a change of technology
The DREX pilot was shut down and the market read it as a verdict on blockchain. It was a verdict on a requirement specification — and in a central registry the same defect reappears in the query, not in the data.
Publications
Original documents, with the premises and the numbers I signed at the time.
-
A Proposed Solution to the Privacy, Scalability and Programmability Challenges of Adopting a Regulated Liability Network (RLN) on blockchain
With Henrique Gomes de Moura (UnB). Introduces the enable address oracle (EAO), the chimeric ledger, the 100 TPS proof of concept and the demonstrated record of 1,000 TPS. Original PDF.
-
Bywise: The next generation of cryptocurrencies as a payment method
With Vitor Sulzbach. The seven pillars, Uniform Data Distribution and the blackboard math of the theoretical 17 million TPS. A document of its time — read it with the label above in mind. Original PDF.
Code
The half of the story that anyone can read line by line.
-
Bywise — fullnode
Consensus, a JavaScript smart-contract virtual machine, an immutability anchor and the slice architecture. A blockchain built from scratch, open, live.
Contact
Open to conversations about digital-asset architecture, applied privacy and settlement.
- Email [email protected]
- LinkedIn /in/felipemarts
- GitHub felipemarts
- Location Brasília, Brazil — available for remote