Most of my work is on Instagram.

Prediction market arbitrage about to become a lot harder!

That is why models like this matter. Instead of assuming every feature carries signal all the time, a transformer can shift weight across returns, volatility and autocorrelation as the data changes. Sometimes that shift is meaningful. Sometimes it is just reacting to noise. The real work is testing whether the attention patterns hold up out of sample or whether the model is just fitting structure that disappears the moment regime changes. #MachineLearning #QuantFinance #Transformers #MarketData #SystematicTrading

most of the work in financial machine learning is getting the data into a usable state if the timestamps are off or the series are inconsistent the research stops being reliable and the model output stops meaning much #WalkForwardValidation #CleanData #MachineLearning #Quant #QuantFinance

HFT economics are driven by scale. If the edge on each trade is small but the number of opportunities is large, lowering latency becomes worth an enormous amount of money. #HFT #QuantFinance #AlgorithmicTrading #MarketMicrostructure #Trading

No cloud credits, no rented minutes. We self host the training stack and wire it straight into the product: 30 live models with tunable parameters, built in feature engineering, and clean datasets ready to run. Open access for anyone who wants to train. #quant #machinelearning #nvidia #dataengineering #selfhosted

Every quant interview gets to conditional probability at some point. In financial markets, probabilities are not fixed. New information arrives and the conditional distribution changes. A large part of quantitative finance is estimating those shifts correctly. That is also why Hidden Markov Models matter. The regime is not directly observed. You estimate the probability of being in each hidden state from observed data, then update that estimate as new observations arrive. #quant #bayestheorem #cardcounting #statistics #quantfinance

Encouraging a 13 year old to start gambling should be a crime.

We’re going to become the best trading platform out there. £0 ad spend.

You can still be a prodigy at 13. Just learn probability and statistics instead of random indicators. #quant

Every trader once thought they could do this 😭

Simulate as many strategies as you want for free Fuck capitalism.

Gold dropped 10% in 24 hours turns out the hedge needed a hedge

LLMs can't tell you what Renaissance does because quant funds spend hundreds of millions on infrastructure and research specifically to know things nobody else knows. A language model trained on publicly scraped text is by construction trained on everything except the stuff that actually generates returns. #llm #quantfinance #renaissance #citadel

This is what you’re competing against; drawing lines on a chart will not give you an edge 😂

Bid and ask are only the first line of the order book. Level 3 data records every order added, cancelled and modified in the book. From that you can reconstruct queue position, liquidity imbalance and how depth shifts across price levels through time. That is the layer most quantitative execution and market microstructure models operate on. #quant #orderbook #level3data #marketmicrostructure #orderflow

The glamorous side of running market infrastructure.

Name an industry Ai cannot steal. You cannot

Companies are paying monthly for tools they could replace with one competent engineer.

The distribution that makes prop firm challenges exploitable. 100k FTMO challenge costs $620. Buy 6 accounts Total cost $3,720 Phase 1 10% profit target 10% max loss 3 / 6 reach phase 2 Phase 2 5% profit target 10% max loss 2 / 6 pass Funded stage 1 account hits 10% drawdown 1 account hits 10% profit Profit on $100,000 at 10% = $10,000 80% split → $8,000 payout $620 challenge fee refund Total received $8,620 Net after $3,720 cost → $4,900 Obviously there are spreads, time limits and rule constraints, but the distribution still leaves the odds favourable if you treat the challenge as a probability problem rather than a trading one. Made over £200,000 from this before turning 18. The method continued to work largely because most traders do not manage risk and rarely think in expected value.

Going to make level 3 market data available everyone; too much gate keeping going on.