I bridge the gap between institutional financial rigor and cutting-edge AI. With over 8 years of SQL mastery and a background in auditing, I build data systems that don't just process information—they map the financial ecosystem.
Building a production-ready platform to map relationships between Hedge Funds, Private Placements (Form D), and Service Providers.
- The Stack: Python (
dlt) → BigQuery →dbt→ Neo4j (Cypher) → Next.js. - The Goal: An AI-driven recommendation engine for fund managers to find the best-fit auditors and custodians based on AUM and fund type.
- View Project Repo | Live Demo (Coming Soon)
- SEC 10-K RAG Pipeline: An end-to-end LLM project using Retrieval-Augmented Generation to query and extract insights from 10-K.
- S&P 500 Predictor (MLOps): A production-grade MLOps pipeline for stock price forecasting. Includes experiment tracking with MLflow, orchestration, and drift monitoring.
- SEC Ingestion Engine: Automated pipeline for large-scale ingestion of 10-K filings using Python and cloud-native tools.
- MLB Win Probability: Machine learning model predicting regular-season game outcomes based on historical performance and real-time variables.
- Car Angle Classifier: A Computer Vision project using PyTorch and transfer learning (MobileNet) to classify car images by viewing angle (Front, Back, Side).
- Experimental projects leveraging AI agents to build functional apps at high speed, focusing on UX and rapid iteration.
| Category | Tools |
|---|---|
| Languages | Python, SQL (BigQuery/Postgres), Cypher (Neo4j), JavaScript |
| Data Engineering | dbt, dlt, BigQuery, Supabase, Segment |
| Machine Learning | PyTorch, XGBoost, MLflow, Scikit-learn |
| AI & LLMs | RAG, Gemini API, Coding Agents (Antigravity), Lovable |
| Reporting/BI | GA4, HubSpot, Metabase, Google Looker Studio |
- LinkedIn: LinkedInProfile
- Email: xchencws@gmail.com
- Location: Remote / Chicago, IL (Greater Area)