A collection of software projects spanning quantitative finance, AI tooling, and various client work. The core focus is wf/ (Wayy Finance) - a suite of Python libraries and a web platform for market analysis and algorithmic trading.
The main event. Five interconnected Python packages that handle everything from pulling market data to deploying live trading strategies.
Universal market data aggregator. One interface, 32+ data providers.
- Free sources: Yahoo Finance, Coinbase, CoinGecko, Binance, Kraken, FRED
- API-required: Alpha Vantage, Finnhub, Alpaca, Polygon.io, Interactive Brokers
- Crypto: Full CCXT integration (100+ exchanges)
- Real-time: WebSocket streaming for Binance, Coinbase, Finnhub, Alpaca, Kraken, Polygon, IBKR
- Extras: Whale transaction detection, Level2 orderbook data
Returns Polars DataFrames. Auto-detects asset type from symbol. Handles the annoying parts of dealing with financial APIs so you don't have to.
Forecasting library based on fractal geometry and chaos theory. Not your typical ARIMA wrapper.
- Hurst exponent analysis - Detect if a market is mean-reverting (H < 0.5) or trending (H > 0.5)
- Fractal dimension - Measure price movement complexity
- Monte Carlo simulation - Generate probability-weighted price paths
- Exogenous predictors - Incorporate VIX, bonds, commodities with automatic lag selection
- Model zoo: ARIMA, GARCH, Prophet, VAR, LSTM, Random Forest, XGBoost, SVR, plus Bayesian models via PyMC
The core idea: markets exhibit self-similarity across timeframes. This library exploits that.
Interactive charting with Polars support. TradingView-style but in Python.
- Standard charts: candlestick, line, area, volume
- Non-standard: Renko, Kagi, Point & Figure, Line Break, Range Bars, Heikin-Ashi
- GPU-accelerated WebGL rendering for tick data (millions of points at 60fps)
- LTTB decimation for high-frequency data
- Jupyter notebook integration
Backtesting and live trading framework. Polars-based, so it's fast (10-50x faster than pandas alternatives).
The workflow:
- Research - Develop strategies, run backtests
- Validate - Permutation testing to catch overfitting
- Optimize - Kelly criterion for position sizing
- Paper trade - Deploy to Alpaca paper trading
- Go live - Deploy with confidence
Key features:
- Component architecture for modular strategy design
- Walk-forward validation
- Hierarchical portfolio management
- CLI for deployment with auto-restart
- 113 tests passing
Full-stack platform tying it all together. FastAPI backend + React frontend (in development).
- REST API exposing wrdata, fracTime, wrtrade functionality
- WebSocket streaming for real-time updates
- Supabase for auth and persistence
- Redis caching
- Claude/GPT integration for analysis
Current status: V2 in active development.
- docs/ - Architecture docs, roadmap, Bloomberg Terminal rebuild notes
- examples/ - Jupyter notebooks demonstrating workflows
- projects/ - Active research (crypto orderflow analysis, breakout detection, ML pipelines)
- wayyFunds/ - Fund prospectus and brand docs
Testing framework for LLM-based agents. Because "it seemed to work" isn't a test strategy.
- Semantic evaluation using Mistral and Claude
- Dashboard for viewing results
- Python package for integration
Vim/Neovim AI code completion plugin. Uses Claude API. Works in both editors (Python for Vim, Lua for Neovim).
Generate Anki flashcards from text or web pages using Anthropic API. Study smarter.
Educational platform teaching AI development through interactive tutorials. FastAPI + React.
AI-powered marketing automation. Multi-tenant platform for:
- Generating ad creative
- Deploying to Google Ads and Meta
- Multi-Armed Bandit optimization
Web archiving with smart content extraction. Firecrawl/Puppeteer for scraping, Gemini embeddings for RAG search, Supabase backend.
Spiritual texts RAG system. Vector search + Neo4j graphs + Claude. Deployed on Render.
Buffalo mayoral voter analysis. Predictive modeling with H3 hexagon mapping and Streamlit dashboards.
Leaf removal business management. React frontend, FastAPI backend, Twilio voice agent, Google Calendar sync. Practical automation for a local business.
Western New York AI community website. React SPA with Three.js, Airtable backend for events.
Mental health companion app (Wayvy). In development.
Windows 95 simulator in React. Draggable windows, taskbar, media player. Pure nostalgia.
Retro 1995-themed social media management platform. Three parts: React UI, Python API, Python CLI.
Full-stack app template. Express/FastAPI + React + Radix UI + Drizzle ORM + Supabase.
Vite + React + Three.js + Supabase starter template.
Figma-to-GitBook sync. Auto-generates documentation from Figma designs.
Quote management and job tracking. Scan items, generate quotes, process payments.
old/ contains legacy projects: bq-ingest, futures-research, rl-fin (reinforcement learning for finance), trading-bot, vol-dashboard, and other experiments.
The common threads across projects:
Languages: Python 3.10+, TypeScript/JavaScript Data: Polars, Pandas, NumPy, PyArrow ML/Stats: scikit-learn, XGBoost, PyMC, statsmodels, ARCH Web: FastAPI, React, Vite, Supabase AI: Claude API (Anthropic), OpenAI, Mistral Infra: Render, Redis, PostgreSQL
A software shop building tools for quantitative finance and AI applications. The core product is the Wayy Finance stack - a modern alternative to legacy trading infrastructure. Everything else supports that mission or pays the bills while we build it.
The philosophy: build useful things, validate them rigorously, ship when ready. No hype, just working software.