Nikita Aull Nikita Aull

Projects

In Development

Lerner

In development
Flutter Dart SQLite Riverpod Go

A mobile and desktop application for learning German, designed specifically for Russian speakers. Covers levels from A0 to B2 with a structured curriculum of 96 lessons.

Key Features

  • Spaced repetition (SM-2) - vocabulary review with independent tracking for words, irregular verbs, articles, and conjugations
  • 4 exercise types - multiple choice, text input, word order, and full sentence construction with progressive difficulty
  • Error loop mechanic - mistakes trigger re-drilling from the error point, ensuring solid learning
  • Built-in dictionary - browse, search, filter by part of speech, add words to review
  • Text-to-speech - native pronunciation for all vocabulary
  • Thematic situations - real-world scenarios: shopping, doctor, bank, government office
  • Offline-first - all content stored locally, no internet required
  • Dark/light theme

Architecture

The app uses a structured 5-stage lesson flow: theory, word learning, grammar check, sentence drills (100 exercises per session), and spaced repetition review.

BugH

In development
Go PostgreSQL SQLite Cobra Rod

A comprehensive bug bounty hunting toolkit combining an automated reconnaissance/testing CLI tool (Scalpel) with a structured knowledge base of 750+ security research files.

Scalpel - CLI Tool

A Go-based command-line tool with 13 core commands for the full bug bounty workflow:

  • Scope - automated reconnaissance with 25 passive sources (crt.sh, DNS, Wayback Machine, Shodan, GitHub, Common Crawl) and 39 active collectors (JS parsing, tech detection, form discovery, GraphQL, Swagger, source maps). Supports stealth mode, resume, and snapshot diff
  • Race - race condition testing engine with 6 send modes (HTTP/2 framing, HTTP/1.1 last-byte, pipeline, HTTP/3) and 9 pre-built recipes (coupon double-spend, cache poisoning, duplicate registrations)
  • Probe - mutation-based anomaly detection with baseline comparison and 13 mutation dictionaries
  • Params - hidden parameter discovery using batch + binary search approach
  • Hunt - AI-powered attack planning with hybrid search (BM25 + cosine similarity + RRF fusion) and LLM-guided questioning
  • CVE Check - local CVE database matching with NVD, EPSS, KEV, and PoC data
  • Nuclei - selective vulnerability scanning based on detected technologies (Wappalyzer DB with 7,360+ tech signatures)
  • BB Watch - bug bounty platform monitoring (HackerOne, Bugcrowd, Intigriti, YesWeHack)
  • Proxy - MITM proxy with YAML-based match & replace rules

Knowledge Base

A collection of 750+ Markdown files organized into categories:

Signal Emitter

In development
Go WebSocket Docker Telegram API

A high-performance real-time market monitoring service, detecting anomalous trading patterns across 7 cryptocurrency exchanges and sending alerts via Telegram. Rewritten from Python to Go, achieving a 1000x speed improvement.

Supported Exchanges

Binance, Bybit, OKX, Kraken, KuCoin, Bitget, Upbit. Spot and futures markets with order book support.

Signal Categories

Detectors analyze changes in market activity:

  • Volume anomalies
  • Price volatility patterns
  • Order book density

Architecture

  • Zero-Alloc architecture
  • Lock-Free architecture

Signal Backtester

In development
Go Plotly.js SSE

An interactive web tool for backtesting market anomaly detectors on historical cryptocurrency trade data. Shares the detector engine with Signal Emitter.

Features

  • Interactive UI with Plotly.js charts for each detected pattern
  • Pattern labeling system for strategy evaluation
  • Multi-worker parallel processing across symbols
  • Automatic download of historical trade data from exchanges
  • Run management with resume support and parameter hashing
  • CSV export of labeled results
  • Real-time progress via Server-Sent Events
  • Fine-tuning detectors for production use

Released

UTIO

Released
Python aiohttp aiogram Telethon WebSocket

A desktop application that automates opening trading instruments in the TigerTrade terminal based on signals from multiple sources. Acts as middleware between trading signal providers and the trading platform.

Signal Sources

  • Telegram Bot - monitors channels and chats via Bot API
  • Telegram Userbot - user-account based monitoring with session persistence
  • WebSocket Client - real-time signal streaming
  • TradingView - webhook integration

Signal Processing

Three parsers work in cascade to extract trading data from any message format: JSON parser for structured data, regex parser for free-form text, and template parser for custom patterns.

Spot Signal Emitter

Released
Python WebSocket Docker pandas matplotlib

A real-time anomaly detection system that monitors trading activity across 6 cryptocurrency exchanges and sends alerts via Telegram when significant price movements or volume anomalies are detected.

Supported Exchanges

Binance, Bybit, Bitget, OKX, KuCoin, Kraken. Both spot and futures markets.

Detection Algorithms

  • Price movement
  • Volume anomalies
  • Density-based detection

Architecture

Multi-threaded pipeline: WebSocket data ingestion, trade processing, DataFrame storage with multi-index, parallel detector analysis, and Telegram alerting with mplfinance charts.

BTC volatility filter suppresses noise during high-volatility periods.

Trading Dashboard

Released
Python Dash Plotly ClickHouse Docker

An interactive web dashboard for analyzing cryptocurrency market microstructure. Connects to a ClickHouse database to visualize trades, order book state, and volume profiles with millisecond precision.

Visualizations

Multi-panel layout with three synchronized views:

  • Volume profile - cumulative volume distribution across price levels
  • Price chart - candlestick line with trade markers (buys/sells) and volume bars
  • Order book - bids and asks with cumulative depth curves at any point in time

Features

  • Time navigation with 100ms granularity
  • Configurable order book depth (0.1% to 100%)
  • In-memory caching of database queries
  • Order book reconstruction from snapshots and incremental updates
  • CSV import for batch analysis of signals
  • Multi-instance deployment (3 parallel dashboards)

Architecture

Built with Dash (Plotly) and ClickHouse as the time-series backend. Dockerized with gunicorn for production.

Crypto Data Analysis

Released
Python pandas scikit-learn mplfinance ClickHouse

A quantitative analysis system for cryptocurrency markets focused on detecting trading anomalies from Binance data.

Analysis Modules

  • Walls detector - identifies large orders and imbalances in the order book
  • Strike detector - detects rapid price spikes (>8% in milliseconds)
  • Feature extraction - linear regression analysis, price noise and fluctuation metrics
  • Imbalance analysis - order book bid/ask imbalance detection
  • Support levels - automatic identification of significant price levels

Data Pipeline

Downloads OHLC and aggregated trade data from Binance REST API. Multi-threaded processing across 1m, 5m, 1h timeframes. ClickHouse integration for large-scale historical analysis.

Orderbook Analyze

Released
Python pandas ClickHouse numpy

A market data analysis system for detecting anomalies in cryptocurrency orderbooks and price movements. Uses ClickHouse as the backend for historical trade and orderbook data from Binance.

Detection Modules

  • Price anomalies - sharp price changes exceeding configurable thresholds with volume concentration analysis
  • Large spreads - abnormally wide bid-ask spreads
  • Sparse orderbook regions - areas with insufficient liquidity and depth gaps

Analysis Pipeline

Trades are grouped by symbol, price, and time window, then filtered by trade count and price change thresholds. Volumes are normalized to USDT using historical base currency prices. Results are exported to CSV with detailed metrics: price change, volume, trade count, volume position.

Binance News Parser

Released
Python Selenium BeautifulSoup requests

A high-speed monitor for detecting new cryptocurrency listing announcements on Binance. Sends instant Telegram alerts when a target token is announced for Binance Futures launch.

Scraping Strategies

Two parallel approaches for maximum speed:

  • Selenium - headless Chrome with optimized settings (no images, disk cache) for JavaScript-rendered pages
  • API requests - direct queries to Binance CMS API with cache bypass techniques (varying page sizes, encoding options)

Features

  • Monitors 100+ target cryptocurrencies
  • Telegram notifications on match
  • Cache bypass via CloudFront header manipulation
  • Process pool with 10 parallel instances for redundancy
  • Timed polling synchronized to specific intervals