Big-data pipeline detecting wash trading and pump & dump in trade data using Apache Spark & Kafka, deployed via Streamlit.
SPK II-15.1 insider-transaction disclosure analytics for BIST - async Python, PostgreSQL, empirical forward-return base rates.
Economic Android watcher agent built on top of PicoClaw: a 24/7 phone‑resident AI that automates apps and web browsing, collects and organizes market/news data, learns over time, and sends structured opportunity reports via Telegram.
🛡️ Physics-inspired RegTech toolkit for financial market surveillance. Detects Pump-and-Dump schemes, Wash Trading, and volume-price divergence using price kinematics.
Detect collusive market manipulation with graph neural networks. NSE-calibrated synthetic-data generator with ABIDES integration + GraphSAGE and Tier-2 GBM detector stack. MTech dissertation work.
Open-source HKEX CCASS shareholding scanner for Hong Kong IPOs. 132 stocks, 109K holder records, 46 empirical findings. Daily automated collection via GitHub Actions.
AI-powered trading intelligence platform combining a 4-tier finance RAG analyst with 24/7 cross-market anomaly detection across CME, Polymarket, Hyperliquid, X, and Truth Social, evaluating the market impact of high-severity signals through comparison with historical market-moving events.
Production-ready crypto market surveillance dashboard for wash trading, spoofing, layering, pump-and-dump, volume spikes, and cross-exchange manipulation.
Deployed crypto market surveillance dashboard with real-time price/volume anomaly detection, wash trading & spoofing pattern detection, alert triage, and PDF case report generation.
End-to-End Python implementation of Zwydak et al.'s (2026) non-parametric framework for flagging artificial trading activity on centralised crypto exchanges. Uses MFDFA multifractal spectra, sign-preserving DCCA, ApEn/SampEn and mean-shift change-point detection on 1-minute tick panels.
Real-time market-manipulation surveillance over live L2 order books — Rust · Kafka · ClickHouse · gRPC · Go · Kubernetes
A real-time crypto market-surveillance pipeline whose analyst model structurally cannot state a number it has not verified. Kraken spot feed, online anomaly detection, explanations checked by a deterministic verifier.
Simulated event-contract market surveillance: SQL + Python detectors for 5 abuse typologies, scored against ground truth.
AI-powered market surveillance agent that monitors Polymarket for insider trading
Market-manipulation detection for ETH/BTC exchange data: Kyle's λ price-impact, round-trip/wash-trade rate, and size-concentration bot fingerprints, benchmarked against cross-venue baselines (Kraken/Binance). Python.
Real-time market-manipulation surveillance over live L2 order books — Rust · Kafka · ClickHouse · gRPC · Go · Kubernetes
Memory-driven market-surveillance agent that detects NFT wash-trading rings and gets sharper with every confirmed case — CockroachDB vector memory + Managed MCP Server, AWS Bedrock
합성 데이터에 시세조종 패턴을 주입해 ground truth를 확보, 정밀도·재현율로 탐지 모델을 엄밀하게 평가하는 시장감시 엔진
🚨 Market-manipulation & RMT surveillance for a live exchange — 6-detector anomaly ensemble, weak supervision, walk-forward eval (AUC 0.93, P@10 1.00), FastAPI + Streamlit
Regulator-side reference PoC for independently verifying anchored Merkle proofs of multi-entity trading logs (VCP Gold/Platinum). Includes Merkle inclusion + external anchor verification and a red/green integrity dashboard.
To associate your repository with the market-surveillance topic, visit your repo's landing page and select "manage topics."