Desmond
Kao
Building AI systems that solve tangible business problems across NYC's financial and tech landscape — from hedge fund data platforms to LLM-powered research tools. I turn messy data into production-ready solutions.
About Me
Originally from San Francisco, I'm a Computer Science & Data Science student at New York University building AI systems that solve tangible business problems across NYC's financial and tech landscape.
I've engineered full-stack platforms and ML pipelines for NYC hedge funds, built automated pricing models for real estate companies, and developed NLP tools for academic research centers including NYU's Carter Journalism Institute and Yale.
Beyond engineering, I'm a classical pianist with 16 years of experience and explore creativity through Muay Thai and music production.
Experience
- Incoming Software Engineer Intern
- Built AI chatbot in Python/FastAPI with Snowflake RAG + OpenAI API, significantly reducing research time for investment staff
- Developed AI-driven commentary pipeline automating daily report generation, synthesizing news, research, and portfolio data
- Created full-stack investor relations platform in React/Python automating chart generation, analysis reports, and ad hoc queries
- Developed firmwide Flask/SQL research API enabling teams to query live data with dramatically improved retrieval speeds
- Built autonomous LLM scraping pipeline in Flask/Azure with daily data collection, significantly reducing research time
- Created AI Excel assistant in Python using Perplexity + Tavily APIs, automating large-scale data entry tasks
- Developed ML valuation API in Python, significantly improving property pricing accuracy and reducing appraisal costs
- Automated end-to-end valuation pipeline from data ingestion to prediction, dramatically reducing review time
- Integrated live market/listing APIs to enrich model features, improving valuation precision across portfolios
- Built LLM/NLP pipeline analyzing large-scale social media posts for sentiment trends using HuggingFace
- Developed large-scale document scraper with LLM-based summarization, significantly reducing publication review time
- Created automated bias detection system evaluating news articles for neutrality and language framing
- Built lightweight LLM with TensorFlow/Keras using pruning and quantization for low-power on-device inference
- Benchmarked model latency and power consumption across CPU/GPU setups for sustainable local deployment
- Led ML/AI sustainability workshop for industry professionals on efficient model design techniques
Projects
Equities-focused AI dashboard integrating Grok API for factor report generation. Statistical analysis engine using SciPy for quant metrics and portfolio analytics.
Firmwide AI chatbot connecting directly to Snowflake for flexible querying across company data. Natural language access to research reports and market data.
Full-stack IR tool for NYC hedge fund connected to Snowflake. Automated data validation pipelines and client-ready report generation.
AI-powered email generator using Pinecone vector DB and OpenAI API. Semantic search across templates with real-time Streamlit dashboard.
Hybrid LSTM–ARIMA model for urban traffic forecasting. Interactive React dashboard visualizing congestion trends as retail/transport indicators.
Custom lightweight LLM built from scratch with TensorFlow/Keras. Sustainable AI via model pruning, quantization, and efficient training strategies.
ML models predicting stock movements after short seller report releases. NLP feature engineering pipeline with logistic regression and deep Q-learning.
Published iOS app for AI-driven dream interpretation and journaling. Firebase encrypted backend with SwiftUI/TCA architecture.
Enterprise document intelligence using Claude API and RAG for natural language querying across company knowledge bases.
ETL monitoring pipeline detecting anomalies across millions of records daily with PowerBI dashboard and LLM-generated summaries.
NLP API extracting structured data from unstructured sources using Gemini and Claude APIs. Containerized with Docker for scalable deployment.
Talent acquisition dashboard using AI agents and blockchain for credential verification. Intelligent matching system for streamlined hiring.
Education
