01Platform
Alpha Lab is my modular trading research platform, focused on the path from market data to strategy evaluation. The work spans data acquisition, strategy and indicator tooling, and custom backtesting, with backtest-to-live architecture under development.
Problem solved
Historical data arrives from different providers in different shapes. Research needs one consistent data contract before any result built on it can be trusted.
Approach
Tech stack
02Research
Four ways a market question becomes testable work. Each answers the same five questions, so they compare side by side.
Study price, trend, momentum, and volatility to turn market hypotheses into testable signals.
01 · Input
OHLCV and price data available at decision time.
02 · Method
Trend, momentum, volatility and mean reversion; for example returns, rolling z-score, ATR.
03 · Output
Explicit features and signal rules with versioned parameters.
04 · Validation
Baseline comparison, out-of-sample testing, sensitivity, regime analysis, transaction costs.
05 · Limits
Parameters overfit. Patterns shift across regimes. Every feature must be causal.
03Selected systems
Eight case studies across custody, signing, contracts, quant research, machine learning, execution, and mining operations.
Coverage
Notable works
Custodial wallet service for multi-chain deposits, pooled wallets, and outgoing transfers, built with stable APIs and layered signing controls.
Evidence
ArchitectureSystem diagram
Approach
The domain core stays apart from the systems it talks to, and the boundary between them decides where a change belongs. That matters most when an LLM helps write the change.
Inbound adapters · call the core
Data provider
Broker / exchange
Database / storage
Outbound adapters · implement a port
Outbound port
Market Data
What the core needs: bars and quotes for an instrument. Data provider adapters implement it: Twelve Data, Bento and Binance. One is wired in at a time; swapping it changes one adapter, while the port and the core stay as they are.
Before an LLM drafts a change, the task is scoped to one side of a boundary: one adapter, or one port and the adapters behind it. That limits the blast radius, gives the implementation a clear target, and makes anything outside the scope stand out in review.
A port is a contract at the boundary: what the core needs from the outside, such as market data or order submission, or what it offers, such as commands. An adapter connects one integration to a port. Outbound adapters, such as a data provider, a broker or exchange, or a database, implement the port the core calls, and several can implement the same port; inbound adapters, such as an HTTP API, a CLI or an MCP server, call a port the core offers. A new provider means a new adapter; new behaviour starts as a change to a port.
Boundaries narrow what a change can touch. They do not make it safe or correct: contract tests, review and validation still decide whether it ships.
04Products · ChartSentry
Indicators built with the same discipline and sold under my product brand. Every object is stamped on a closed bar and never rewritten.
TrueSMC
A five-module smart-money suite for MT5: market structure, order blocks, fair value gaps, liquidity and premium/discount zones from one engine.
StackMQL5 on MetaTrader 5
TrueSMC on chartsentry.net (opens in a new tab)TrueSMC Lite, free (opens in a new tab)
Pattern Sentinel
The core 12 classical chart patterns, quality-scored, with every resolved call kept on the chart as a permanent track record.
StackMQL5 on MetaTrader 5
05About & contact
I work at the intersection of trading and software engineering. I have traded hands-on since 2007, so the systems I build start from how a strategy actually runs: what data exists at decision time, what costs and fills look like, and what fails first in a live session.
I have also worked as a backend engineer at Jagad, since 2024, on smart contracts and swap APIs, and at Tokenomy, from 2021 to 2023, on node middleware, wallet operations and asset accounting for a crypto exchange. In 2019 I built neural-network and reinforcement-learning systems for trading on contract.
Focus
Trading systems first: market data, strategy tooling, backtesting and automation, with Alpha Lab as my current platform. Markets traded hands-on: forex, commodities, equities and crypto.
Work that fits
Two kinds of engagement, for traders, teams and funds.
Engagement 01
End to end
From market data to execution: data contracts, strategy logic, backtesting with realistic costs, and a measured path to live trading.
Typical shape
Engagement 02
Your rules, automated
Turn a rule-based or discretionary strategy into a system that behaves the same way every session: rules made explicit, tested out of sample, then automated.
Typical shape
Contact
LinkedIn (opens in a new tab)GitHub (opens in a new tab)
Hiring for a trading systems or core platform role? Skip the booking: email or LinkedIn reaches me directly.
Consultation
We cover your strategy or system, the data it needs, and what a first milestone looks like.