Ajit Nath

  • Mobile: +852 53691903
  • Location: Hong Kong

Professional Summary

Software Professional with a background in data science and engineering. Passionate about developing and deploying software solutions to solve real-world business problems.

Professional Experience

Glassbox AI (Part-Time)

  • Period: 2025 – present
  • Summary:
    • Working on building a sign language platform that generates Hong Kong Sign Language videos from Traditional Chinese transcripts
    • Working on developing Office Add-Ins for the Hong Kong Monetary Authority leveraging their in-house inference platform

Ubizense Limited

  • Period: 2022 – 2025
  • Summary:
    • Worked on data analytics projects at Hong Kong International Airport, including:
      • Used computer vision techniques to monitor baggage carousels for arriving bags
      • Conducted data analysis and built statistical models to predict flight departure delays
      • Fine-tuned LLMs to generate structured data from free-text logs for classifier input
      • Developed an ETL pipeline for airport vehicle GPS data

Oriental Patron Investment Management Limited

  • Period: 2017 – 2022
  • Summary:
    • Built a trading platform using Python, Bloomberg API, and PostgreSQL
    • Processed raw data from the Bloomberg API to generate reports for buy-side fundamental analysis
    • Periodically scraped financial news and applied BERT for generating embeddings to identify relevant articles for market analysis

Tofugear

  • Period: 2015 – 2016
  • Summary:
    • Developed web applications using ReactJS, Ruby on Rails, and PostgreSQL
    • Focused primarily on frontend solutions for SMEs

City University of Hong Kong

  • Period: 2014 – 2015
  • Summary:
    • Worked as a research assistant under Prof. Samson Young
    • Contributed to research on developing novel electronic musical instruments for music therapy in autism

Technical Skills

  • Programming Languages and Frameworks: Python, SQL, JavaScript, Rust, C++, FastAPI, Flask, PyTorch, statsmodels, Scikit-learn, XGBoost, OpenCV
  • Machine Learning Techniques:
    • Regression and time series analysis using statsmodels
    • Tabular data-based ML using Scikit-learn
    • Computer vision using YOLO for object detection, SAM for image segmentation, and OpenCV for general-purpose image data processing
    • NLP using large language models, distillation, and fine-tuning LLMs using LoRA and QLoRA
    • Text similarity analysis using embeddings
  • Data Engineering:
    • ETL/ELT pipelines
    • Big data analysis using GCP
    • Processing large file-based datasets using DuckDB
    • Analyzing large geospatial datasets with Polars, Pandas, and GeoPandas
  • Development Tools and Practices: Git, Docker, Agile methodologies, Test-Driven Development

Education

  • Master’s in Computer Science | City University of Hong Kong | 2014
    Project: Improvisor: Genetic-algorithm-based music improvisation github
  • Bachelor’s in Computer Science | Himachal Pradesh University, India | 2012

Publications

  • Vesball: A Ball-Shaped Instrument for Music Therapy
    Link
  • Activity Tracking and Monitoring of Patients with Alzheimer’s Disease
    Link

Personal Projects

  • CHIP-8 Game Emulator (Rust, WASM)
    GitHub
  • Simple-RTMW (Easy-to-use implementation of RTMW pose detection inference)
    GitHub
  • KMB-ETA (Open-source KMB bus tracking web app)
    GitHub
  • Other open-source contributions to larger projects such as polars, a fast, efficient DataFrame library written in Rust.