I'm a Computer Science PhD specializing in audio processing, multimodal retrieval, and machine learning systems.
These days, I enjoy building AI-powered Telegram bots, working on quantitative trading tools, and exploring computer vision + LLM agent workflows.

I also love photography (Sony A6700), gaming, and flying my drone.
This GitHub is a mix of research code, practical ML pipelines, and personal side-projects

• GitHub: https://github.com/dhecloud
• Email: andrewkjj@gmail.com
• 📄 Check out my CV / Resume

Ph.D. Computer Science – Nanyang Technological University (2020–2024)
Thesis: Audio Captioning and Retrieval with Improved Cross-Modal Objectives

Published research in:

• Automated Audio Captioning
• Language-based Audio Retrieval
• Acoustic Event Detection
• Word Sense Disambiguation (BERT)

See publications below 👇

A context-aware Telegram bot built with LangChain, RAG pipelines, and multilingual embeddings.
Features include:

• JLPT grammar/vocab scraping + JMDICT + Tatoeba + JaQuAD
• Automated Japanese news explanations via multi-agent workflows
• Cloze-question generation from YouTube transcripts
• Memory persistence + metadata filtering improvements
  👉 Bot: https://t.me/SakuraSenseiNoBot

A production-grade Telegram bot that swaps user faces into GIFs, short videos, and “Featured Clips”.
Highlights:

• ONNX-based inference for fast face swapping
• Usage tracking, quotas, and Stripe premium tier
• Watermarking + file size/duration validation
• Migrated webhooks from Ngrok → Cloudflare Tunnel
  👉 Bot: https://t.me/FaceChangerGIFBot

• Language-based Audio Retrieval with Converging Tied Layers and Contrastive Loss (APSIPA 2022)
• Automated Audio Captioning with Epochal Difficult Captions for Curriculum Learning (APSIPA 2022)
• Audio Captioning with Reconstruction Latent Space Regularization (ICASSP 2022)
• Sound Event Detection with Weakified Strong Labels & Frequency Dynamic Convolution (arXiv 2023)
• Adapting BERT for Word Sense Disambiguation with Gloss Selection (EMNLP Findings)

Machine Learning: PyTorch, TensorFlow, ONNX, transformers, LangChain
Domains: Audio Processing, Computer Vision, NLP, Cross-Modal Retrieval
Tools: Docker, Cloudflare Tunnel, Whisper, VAD, BeautifulSoup
Languages: Python, English (Native), Chinese (Conversational), Japanese (Conversational)

Thanks for visiting!
Feel free to explore my projects or reach out if you'd like to collaborate.