One book, one desk
We run a single managed pool instead of a catalogue of products. Scale is what buys institutional pricing and keeps the cut per position small.
About
NeuralCryptoCapital pools client cryptocurrency into a single managed book, then trades it with our own models. Clients choose an asset and an amount; the desk does the rest.
The company was built around a simple observation: most private investors in cryptocurrency are exposed to the worst version of it — late entries, emotional exits and retail pricing. A pooled book fixes the economics, because size changes what a trade costs.
Client capital joins the pool and is allocated across the assets the desk holds. The trading program times entries and exits; the risk desk sets what the program is allowed to do. Your account is valued every day and every movement is written to your history.
The engineering team behind the engine has worked on production machine-learning systems at OpenAI, Google DeepMind and Meta AI, and now builds alongside the research teams at Anthropic and OpenAI. That background is why the desk is documented: models, mathematics and limits are published rather than implied.
NeuralCryptoCapital does not provide financial advice or individual investment recommendations. Performance displayed on this site is simulated for demonstration.
$160M
Assets in the pool
$1,280M
Lifetime investment volume
120%
Target · 12 months
24/7 · automated
Desk
We run a single managed pool instead of a catalogue of products. Scale is what buys institutional pricing and keeps the cut per position small.
Every trade is produced by the trading program and recorded with its inputs. Discretionary overrides are logged like any other decision.
Exposure limits, a loss budget and a kill switch sit ahead of any signal in the pipeline.
Clients see their own allocated capital, yield and daily valuation — not a marketing curve detached from the account.
The founder
Founder & Chief Technology Officer
Palo Alto, California
“The edge is not a signal. It is a network of models trained by people who actually understand the mathematics — and a desk disciplined enough to stop them.”
Marcus built NeuralCryptoCapital after a decade spent on the infrastructure behind frontier machine learning. He holds a Ph.D. in Computer Science from Stanford, specialising in deep learning and matrix equations, and a B.S. in Applied Mathematics and Software Engineering from MIT. His work sits at the intersection of model architecture and the risk controls that make it safe to put capital behind one.
Experience
Founder & Chief Technology Officer
2024 — Present
NeuralCryptoCapital
Built the desk, the agent network and the valuation engine.
Principal Machine Learning Architect
2021 — 2024
OpenAI · Google DeepMind
Led model infrastructure for frontier reasoning systems.
Senior Software Engineer, AI Infrastructure
2017 — 2021
Meta AI
Distributed training and inference at scale.
Education
Ph.D. Computer Science — Deep Learning & Matrix Equations
Stanford University
2017
B.S. Applied Mathematics & Software Engineering
Massachusetts Institute of Technology
2013
SIPC Insured
Verification pending
FinCEN Registered MSB
Verification pending
Data Encryption Secured by Cloudflare SSL
Verification pending
Registration takes a minute. Pick the crypto you want to invest, add it to the pool and follow your account daily — funding, history and withdrawals all in one place.