Everything from an idea to a monitored live system
Fixed scope, fixed price, written deliverables. You keep the code, the data and the broker relationship. We do not take a share of your trading results and we do not quote expected returns.
Strategy development
We take a strategy you already understand — from a document, a spreadsheet, a TradingView script or a conversation — and build it as software that behaves identically in backtest and in production.
- Written specification signed before coding
- Python or C++ implementation in your repository
- Deterministic replay and structured decision logs
- Expiry, holiday, corporate action and lot-size handling
- 30 days of defect fixes after delivery
Backtesting and validation
We rebuild your test so the result survives contact with reality: point-in-time data, full costs, honest fills, walk-forward folds and stress windows.
- Data audit before any result is trusted
- Full cost stack: brokerage, STT, exchange, GST, stamp, impact
- Walk-forward folds and parameter surfaces
- Regime, volatility and crisis-window splits
- Trade-by-trade log handed over with the report
Strategy improvement
Diagnostic work on a live or nearly-live strategy: where the money is actually lost, which rules add value, and which ones you are paying to keep.
- Attribution: where P&L is made and lost
- Rule-by-rule ablation with evidence
- Risk and drawdown control redesign
- Execution and timing improvements
- Every accepted and rejected idea documented
Execution engineering
The layer between a signal and a fill: broker APIs, retries, idempotency, reconciliation, kill switches and measured slippage.
- Broker API integration with reconnect and re-auth
- Idempotent orders and client order IDs
- Slippage measurement and entry-timing tuning
- Position reconciliation against broker statements
- Kill switch, loss caps and order-rate limits
Cloud deployment and hosting
Deployment on AWS, GCP, Azure or a plain VPS, with scheduling, secrets handling, alerting and a runbook your team can follow.
- Containerised deployment on your account
- Secrets management, no credentials in code
- Market-hours scheduling and auto-restart
- Telegram or email alerting with heartbeats
- Backups, log retention and a written runbook
Monitoring and market-hours support
Heartbeats, alerting, daily reconciliation, incident response during market hours and a monthly health report.
- Heartbeat and staleness monitoring
- Alert routing to Telegram, email or phone
- Daily P&L and position reconciliation
- Incident response during market hours
- Monthly health and error-budget report
Code review and independent audit
An independent read of the code, the tests and the claims, delivered as a findings document with severity ratings.
- Correctness, look-ahead and cost-model review
- Risk control and failure-mode assessment
- Security review of credential handling
- Vendor claim verification where evidence exists
- Findings document with severity and fix order
Market data infrastructure
Historical downloaders, incremental coverage ledgers, tick and OHLC storage, corporate action handling and quality checks.
- Historical downloaders with rate-limit handling
- Incremental backfill, no full re-download at startup
- Coverage ledger and gap detection
- Corporate actions, splits and expiry mapping
- Automated data quality checks and alerts
Not sure which one you need?
Most people arrive with a strategy that works on paper and stalls in production, or a backtest they no longer trust. The first call is diagnostic: we look at the rules, the data and the fills, and tell you which of the above actually applies — even if the answer is "none of them, fix your data first".
Book a review call