Skip to content
View kenny1031's full-sized avatar

Block or report kenny1031

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
kenny1031/README.md

Hi, I'm Kenny 👋

CS & Financial Maths @ USYD | ML Systems | Search/Ranking | Backend | Data Science
Building end-to-end ML, analytics, and backend systems.

🚀 Featured Projects

  • Video Delivery Quality & Cost Analytics Platform — End-to-end analytics platform for video delivery quality, CDN cost governance, and edge capacity analysis, combining real YouTube metadata, synthetic CDN delivery logs, PostgreSQL, Python/SQL metric pipelines, A/B testing, Streamlit dashboards, and local LLM-assisted reporting
    Tech: Python, SQL, PostgreSQL, Streamlit, pandas, NumPy, YouTube Data API, Bash, Ollama, A/B Testing, Bootstrap Confidence Intervals

  • Airbnb Retrieval Ranking Engine — Full-stack natural-language Airbnb search system with dense semantic retrieval, structured candidate generation, and an XGBoost learned reranker
    Tech: Python, FastAPI, PostgreSQL, SQLAlchemy, React, TypeScript, XGBoost, Sentence Transformers, Docker

  • Geospatial Recommendation System for Site Selection — Recommendation system that ranks (area, industry) pairs to answer both “what business to open here” and “where to open this business”, using large-scale location data and spatial features
    Tech: Python, PostgreSQL, SQL, XGBoost, PyTorch, FastAPI

  • Regime-Aware Dynamic Asset Allocation — Multi-model quant research pipeline for market regime detection and dynamic asset allocation, combining HMM/GMM/KMeans, XGBoost, LSTM, walk-forward backtesting, and a PPO reinforcement learning prototype
    Tech: Python, pandas, scikit-learn, PyTorch, XGBoost, Stable-Baselines3

  • Helmet Detection System (YOLOv8) — End-to-end computer vision pipeline covering data preprocessing, training, evaluation, error analysis, and inference service deployment
    Tech: Python, PyTorch, YOLOv8, FiftyOne, FastAPI, pytest

🧰 Tech Stack

ML / Retrieval / Analytics: Python, PyTorch, scikit-learn, XGBoost, Sentence Transformers, A/B Testing, Bootstrap Confidence Intervals
Backend / Data: FastAPI, PostgreSQL, SQLAlchemy, SQL, Docker, Git, Bash
Applications: Streamlit, React, TypeScript, FiftyOne, Ollama, YouTube Data API
Also used: C, Java, R

📌 What I'm focusing on now

  • Building production-style ML, analytics, and backend systems
  • Interested in machine learning, search/retrieval, ranking, data science, computer vision, and backend engineering

📫 Contact

Pinned Loading

  1. airbnb-retrieval-ranking-engine airbnb-retrieval-ranking-engine Public

    Python

  2. Checkers Checkers Public

    Java

  3. cv-helmet-detection-system cv-helmet-detection-system Public

    Python

  4. regime-aware-dynamic-asset-allocation regime-aware-dynamic-asset-allocation Public

    Multi-model framework for market regime detection and dynamic asset allocation using HMM, XGBoost, LSTM, backtesting, and RL.

    Python

  5. llm-video-delivery-analytics llm-video-delivery-analytics Public

    A local data analytics project for video delivery quality, CDN cost governance, and edge resource capacity monitoring.

    Python

  6. fintie/smart-developer fintie/smart-developer Public

    Python