Mohammad Muavia is the founder of MIZAN, the trust layer for machine finance, and the inventor of the verifiable track record — to our knowledge the first system that proves a trading record is real, human or AI, without revealing the strategy behind it. His work closes a problem the field's own literature documented, and left open, for forty-six years. He is the author of VTR-1, the open standard for verifiable track records, and of three research papers on SSRN. A builder and trader in digital assets through every cycle since 2017, he has founded or co-founded six ventures, including two exits, and built and shipped MIZAN before raising outside capital.
Most biographies ask to be believed. This one is checkable — which is also what its subject builds. The registry and its 83 credentials: studio.mizan.market · the standard: SSRN 7215664 · a live credential reproduced in ~200 lines of open Python: github.com/muaviamohammed · the live chain's anchor: Bitcoin block 962,013.
Mr. Muavia's first Bitcoin transactions were in 2014 — purchasing developer tools, when that was still the hard way to pay for anything. He has operated in digital assets through every cycle since 2017: the ICO era, the 2018 winter, the 2020 DeFi expansion, the 2021 peak, the 2022 collapse, and the institutional period that followed.
He entered the market as a builder. In 2020 he co-launched Unique.one, one of the earliest NFT marketplaces — establishing a two-sided market in a category that did not yet have a name — and followed it in 2021 with Unique.foto, the first NFT marketplace built specifically for photographers. Both ventures were exited, at 30× and 21× respectively. Alongside, he spent roughly five years as an investor and adviser: early-stage positions, frequently pre-launch, across the Polkadot and TON ecosystems and DeFi's 2020–21 cycle; hands-on research and advisory for dozens of founders on incentive design, market structure and go-to-market — almost all of it done anonymously, under handles rather than a name, which is how that part of the market actually worked and why none of it accrued to a public résumé; and more than $200,000 in lifetime gas fees across his wallets, a record of participation that anyone can inspect on-chain. The one venture that carried his own name in public was Convival.io, a Web3 gaming company backed by Aptos; across his ventures he won multiple ecosystem grants.
He was also, throughout, an active trader: years of proprietary crypto trading and cross-venue arbitrage — spot, perpetuals, funding-rate spreads — with daily working knowledge of fee tiers, slippage, execution quality, margin and liquidation mechanics. This is why MIZAN's gate charges real costs inside the proof: its author spent years on the side of the table where the costs are real.
From 2022 to 2026 he served as Chief Executive Officer of BlockNext Labs, a Dubai-based full-stack crypto studio spanning product engineering, smart contracts, on-chain infrastructure and go-to-market advisory. Across many client teams the same pattern kept repeating, and it became MIZAN's founding premise: in digital assets, anyone can claim traction; what is scarce is proof.
From 2023, Mr. Muavia built ATLAS, an autonomous quantitative research platform, and spent three years building and honestly evaluating more than thirty systematic strategy families across Binance and Hyperliquid — trend, carry, and order-book microstructure — with full execution infrastructure: continuous signal daemons, a real-money live pilot, and multi-venue data recorders across 230 perpetual markets.
The defining output of that period was not a strategy. It was a discipline: more than forty of his own strategies publicly retracted, each with its documented root cause, on a permanent public ledger. The verification gate he subsequently built judged its own author first — it failed his flagship strategy (Deflated Sharpe 0.68 against the 0.95 bar), refused his own S&P 500 strategy on survivorship-free data, and refused thirty-four US equities including the entirety of NVIDIA's price history — the full analysis is public in the survival table. Each verdict is published with its proof on the registry.
He did not set out to build it alone. He first tried to buy it: he self-funded quant and zero-knowledge engineering teams to build the verifier, and they could not — proving a full backtest and its statistics inside a zero-knowledge proof had no blueprint. So he built it himself: twelve engine eras, 18,000+ lines of production Rust, an open standard and three papers.
The input every anti-overfitting correction requires — the number of strategies tried — was self-reported for the twenty-five years the corrections existed. His construction commits the trial set to a Merkle tree before evaluation, proves the count as the leaf count, and recomputes the Deflated Sharpe Ratio against it inside a zero-knowledge proof. To our knowledge, the first mechanism of its kind — specified in full in the mechanism paper.
A strategy hashed and anchored — Bitcoin and RFC-3161 — before the data on which it will be judged exists, then re-proven on the growing window, each link cryptographically bound to its parent. The question "was this built in hindsight?" reduces to a millisecond check. The live chain is public, anchored in Bitcoin block 962,013.
A neural network's complete trading record proven in-circuit with the model's parameters never revealed. His own highest-performing model — +895% — was submitted to the gate he sells: refused on risk, published in full.
Forty-one normative requirements, three conformance levels, nine test vectors, era-law governance, and a public errata register that includes findings against his own implementation, under published governance. The standard is on SSRN under CC BY — free for anyone to implement, including competitors.
The zero-knowledge verification engine, self-serve product, public registry and open standard — built and shipped before raising outside capital. RISC Zero zkVM proving on data-center GPUs, Bitcoin anchoring, era-versioned trust roots; a public estate of forty-plus pages; used, uninvited, from 13 countries and visited from 56, measured in server logs with bots excluded; senior institutional inbound.
Thirty-plus strategy families honestly evaluated; the public retraction ledger that became MIZAN's founding thesis.
Contributed systematic-trading, market-signal and AI expertise through a period of roughly fivefold growth; departed cleanly to focus on MIZAN.
Full-stack crypto studio: engineering, advisory, incubation and growth across the venture lifecycle.
Backed by Aptos; the one venture he built under his own name rather than a handle. Shipped and operated a consumer product through a full market cycle; when the market turned, wound the venture down honestly rather than pivot for appearances.
Among the first marketplaces of its kind — a focused bet on a single underserved creator vertical.
Among the earliest NFT marketplaces — liquidity, creator onboarding and market-making in a category without precedent.
Early-stage investor in blockchain technology innovations — infrastructure, DeFi and digital-ownership ventures, frequently pre-launch; positions across the Polkadot and TON ecosystems and the 2020–21 DeFi cycle. Participation verifiable on-chain since 2017.
Ongoing advisory appointments on product, market structure and go-to-market.
AI-directed engineering, practiced since 2023: architecture and service boundaries, specifications precise enough to implement without guesswork, decomposition into verifiable units, adversarial review and red-teaming of his own releases — reviewer of record for every line that ships. A 128-credential automated regression suite; machine-checked consistency between every published figure and the live registry — the full internal security review is published, including findings against his own implementation. Correctness enforced by machinery, not trust — the method and the product are the same idea.
Mizan Verification, Inc. · Delaware C-Corporation · mizan.market · Standard: VTR-1 (SSRN, CC BY) · Registry: append-only, Bitcoin-anchored · Dubai / remote