vickyonchain.nyc  //  NYC
open to on-chain data & compliance analyst roles — summer 2026

Raw chain data, turned into decision-grade signal.

I'm Vicky — a data analyst in New York with a mathematics degree from NYU Courant. I build fraud detection systems, on-chain intelligence agents, and analytics dashboards. My work sits where raw blockchain data has to become something a risk team, a trading desk, or a compliance analyst can actually act on: validated models, clean pipelines, honest metrics.

B.A. Math NYU Courant Chainalysis + Elliptic certified 4 languages EN / 中文 / FR / IT
Selected work
0x01 ChainSentinelAI compliance copilot — ETHGlobal New York 2026 ◌ in build
An AI copilot that monitors on-chain activity, screens transactions against AML and sanctions risk patterns, and generates plain-English compliance reports through a multi-agent pipeline. Extends my FraudLens codebase; built for ETHGlobal's Continuity Track.
Claude API · agent orchestration · risk scoring · Ethereum
0x02 FraudLensFraud detection analytics platform ✓ shipped
A fraud detection platform mirroring Plaid Protect functionality: engineered features, XGBoost classification with strict time-based splits to eliminate lookahead bias, and a BacktestEngine that validates detection performance against historical data.
Python · XGBoost · time-series validation · backtesting
→ github.com/YOUR_GITHUB/fraudlens
0x03 Onchain IntelligenceThree-agent analysis pipeline ✓ shipped
A Data → Analyst → Report agent system built on the Claude API that ingests Polymarket data and produces structured market intelligence — each agent with a single responsibility, chained into an end-to-end research pipeline.
Claude API · multi-agent design · Polymarket data
→ github.com/YOUR_GITHUB/onchain-intelligence
0x04 Dune Analytics dashboardsSolana ecosystem protocol research ✓ live
Public dashboards profiling Solana ecosystem protocols including Loopscale and Orca — protocol revenue, user behavior, and liquidity flows queried directly from chain data.
SQL · Dune · Solana
→ dune.com/YOUR_DUNE
Background

The through-line

I trained in mathematics at NYU's Courant Institute, and the habit stuck: before I trust a number, I want to know how it was made. That's why my fraud models use strict time-based splits, my probability models get calibrated, and my dashboards query chain data at the source. Whether the question comes from a risk team or a protocol, the job is the same — separate signal from noise, and show your work. I currently work as a data analyst at a New York nonprofit while building in web3.

Credentials & skills

  • Chainalysis Academy certification AML / tracing
  • Elliptic LEARN certification compliance
  • Python, SQL, XGBoost, agent pipelines build
  • Dune Analytics, on-chain forensics analyze