← All projects
AlphaPulse
Algorithmic Trading Engine
Live · Fintech / Systematic Trading · 2025 – Present
Role: Co-builder — system design, signal logic, risk framework
Most retail trading fails for reasons that have nothing to do with market knowledge. It fails because a human sits between the signal and the execution, and that human is tired, anxious, over-confident after a win, or gun-shy after a loss. The edge gets eaten by the emotion.
AlphaPulse was co-built to remove that gap entirely. It is a rule based trading system: signals are generated from explicitly defined conditions, every strategy is validated against historical data before a single rupee is risked, position sizing and risk limits are enforced by the system rather than by willpower, and execution happens automatically once conditions are met.
The design philosophy is deliberately unfashionable. It does not use an opaque machine learning model that produces a number nobody can interrogate. Every signal traces back to a rule you can read, argue with, and disprove. When the system loses money, and it does, you can point at exactly which rule was responsible and decide whether to fix it or kill it.
Outcomes
- Runs live with fully automated execution — no manual intervention between signal and order
- Every trade traceable to a named rule, making post-mortems specific instead of speculative
- Downside bounded by system-enforced risk limits rather than by discretion
What it taught me
- A rule you can read and disprove is worth more than a model you have to trust.
- Backtesting is a gate, not a formality. A strategy that has not survived history has not earned live capital.
- Risk controls, not signal quality, determine whether a system survives long enough for the edge to matter.
- The same loop that runs a good growth program runs a good trading system: hypothesis, test, kill criteria, measure, repeat.
Permalink