Projects /algoforgex
AlgoForgeX
A financial intelligence platform for retail and prosumer investors, built around AI: portfolio and risk analytics, backtesting, on-chain crypto research, decision support, and a multi-tenant advisor and client portal. Built solo with heavy AI assistance, around 420,000 lines of .NET 9 Blazor running live on Azure. Informational and educational, not a brokerage.
AlgoForgeX is a live financial intelligence platform for retail and prosumer investors, built around AI. It is not a brokerage and not a robo-advisor. You cannot place a trade or move money through it. Everything is informational and educational. What it offers is depth: research, analytics, and decision support powered by AI across markets, portfolios, and risk.
What it does
- Market data and research. Detail pages for each instrument, on-chain research for Bitcoin and Ethereum with a classifier for the market cycle, a unified research workbench that runs a question through more than one model, a conversational Ask tab, a way to compare instruments side by side, and a rates and macro view.
- Portfolio and risk. Portfolio tracking with FIFO profit and loss, a risk dashboard (Sharpe, Sortino, value at risk, beta, mean-variance optimization), a backtesting engine with walk-forward analysis, Monte Carlo, and grid search, paper trading, and dividend and bond ladder income tools.
- Decision support, powered by AI. A portfolio assistant that streams a health narrative and draws its own charts, forecasts that reason in several steps, trade ideas with tracked outcomes, and a synthesis board that reads research, ideas, and backtests from across the app for convergence.
- Reflection and calibration. A trade journal that scores how well conviction lines up with execution, and smart watchlists built over more than thirty metrics.
- A chart builder. A charting workbench, built from scratch, where chart definitions are saved as versioned data, so the rendering engine underneath can be swapped without losing a saved chart.
- The business layer. Subscriptions and billing wired to Stripe, an advisor and client portal with white-label branding, and an admin area that tracks the cost of every AI call, broken down by feature.
The build
One developer built AlgoForgeX in roughly six months, mostly with an AI coding assistant. It runs in production on Azure: about 420,000 lines of .NET 9 Blazor, around 580 CQRS handlers, 142 domain entities, 356 database migrations, and roughly 2,850 automated tests, a healthy 1 to 5.5 ratio of tests to the rest of the code.
The throughput is the AI story. The shape is the discipline story. Volume is cheap now; an assistant will generate features faster than you can review them. In fintech, where the product is numbers and a wrong number fails silently, the work that takes judgment is the system built around the AI so that fast does not also mean wrong.
Architecture highlights
- Clean Architecture with CQRS via MediatR.
- Data security in three layers: Azure SQL transparent data encryption, EF Core global query filters that scope every read to the current user, and AES column encryption on sensitive fields.
- A provider abstraction so AI models (Claude, OpenAI) and market data providers can be swapped per feature through configuration.
- Every AI call wrapped in one shared guardrail prompt: informational only, never directive.
- Charting kept independent of any one engine, and a discipline of shipping new features turned off behind flags so releases to a live production app stay safe.
The build series
I am documenting how this was built, honestly, in an ongoing series: Building AlgoForgeX. It covers the day to day workflow, the architecture decisions, the production issues, and an honest account of what the AI did and did not do.
Visit the live product at algoforgex.com.