Engineering Portfolio
Veridian: Autonomous Real Estate Intelligence
An enterprise PropTech pipeline combined with an autonomous, tool-using AI Analyst. It orchestrates the harvesting of unstructured global property listings using headless Playwright automation, normalizes them via a Pandas data warehouse, and stores them via SQLAlchemy. Features a Zero-Framework ReAct loop binding to Groq's Llama 3.3 70B, allowing the agent to route execution between SQL toolkits, live web gateways, and deterministic financial calculators.
Aisle — AI Retail Intelligence Agent
An autonomous AI analyst for e-commerce. Aisle ingests store data into a consolidated warehouse and places a multi-tool AI agent on top. Uses a native tool-calling loop (Groq/Llama 3) with 12 tools to forecast demand, recommend pricing, and segment customers (RFM analysis). Responses stream token-by-token to a premium Next.js dashboard.
Automated Competitor Price Tracker
An end-to-end Data Engineering pipeline that automatically extracts e-commerce pricing data via Python Requests, normalizes and enforces strict typing using Pandas, and loads it into a relational SQLite database via SQLAlchemy. Exposes secure REST API endpoints via FastAPI to serve real-time analytics.
AI Web Scraping Agent (SSRF-Hardened)
An autonomous agent taking plain-English instructions to extract validated JSON from any website. Uses a two-stage fetch architecture to save costs. It dynamically generates extraction schemas at runtime and is fully SSRF-hardened by validating IPs against loopback and cloud-metadata ranges.
MarketSync: Multi-Source ETL Pipeline
Extracts live job market data from isolated sources (REST APIs, Web Scrapers via BeautifulSoup, static CSVs), cleans it with strict Pandas logic, and unifies it into a single PostgreSQL Data Warehouse. Served via FastAPI to a responsive Tailwind dashboard.
NASTP Air-Gapped Examination Network
An enterprise-grade, offline Client-Server architecture for high-security aerospace facilities. Bypasses the cloud by establishing an asynchronous HTTP polling network over a local subnet. Features an automated Pandas ETL engine to ingest messy CSV question banks.
FastAPI Data Ingestion Engine (DaaS)
Enterprise FastAPI backend for secure, high-concurrency asynchronous data processing. Implements strict OAuth2/JWT authentication, Alembic migrations, and Domain-Driven Design (DDD). Uses FastAPI BackgroundTasks and Pandas to process massive CSV workloads instantly.