Hazqeel Afyq

Malaysia

"Curiosity is the wick in the candle of learning."

I believe that curiosity is the wick in the candle of learning. This mindset carves me to become the jack of all trades and master of one (not none). Thus, I have been applying and polishing my skills in various parts while also focusing to master both Python and Machine Learning, especially in Natural Language Processing (NLP), Computer Vision (CV) and Healthcare sector. Furthering my study for a master's degree taught me about researching and how to understand models better. Consequently, helping me to add methods that can improve models' performances.

"Alone we can do so little; together we can do so much" - Helen Keller

I also believe that we as humans can only accomplish so little alone. Hence, I have been actively joining or leading some events during my study to learn and improve my teamwork and communication skills. To substantiate further, I have created two e-books and a website together with my team.


Experiences

Full Stack Developer

Revenue Monster
  • Enhanced an existing customer support chatbot by employing the Electra Small model, fine-tuned through knowledge distillation from Electra Large.
  • Developed an efficient facial recognition enrollment pipeline for biometric attendance that requires only a single image per user and completes preprocessing in under 10 seconds to generate high-quality facial embeddings.
  • Developed a new, more efficient facial liveness model for e-KYC that outperforms current model in both speed and accuracy.
September 2023 - Present

Full Stack AI Developer

Ellzaf
  • Developed and maintained web application and AI using Django, Python, PostgreSQL, PyTorch, JavaScript, HTML, CSS and AWS.
  • Reduced 90% of AI model inference time by replacing DistilBERT with ELECTRA-Small while reduction in model performance was addressed by implementing additional data and text augmentation techniques.
  • Achieved 60% reduction in website loading time by opting for self-hosting assets instead of relying on a content delivery network (CDN).
  • Applied AWS Elastic Load Balancer to improve application scalability.
  • Involved in developing two Python eBooks for beginner and intermediate programmer.
  • Mentored secondary school students about C++ and Python.
  • Delivered Introduction to AI Talk for a group of more than 30 Yayasan Pahang staffs.
  • Delivered Artificial Intelligence Career Talk to an audience of over 50 university students.
  • Delivered Career Talk to an audience of over 50 MRSM PC students.
January 2023 - August 2023

Data Scientist

Aibots MY
  • Researched road safety related issues for Malaysian Institute of Road Safety Research (MIROS) by utilizing the Twitter API to collect tweets from across Malaysia.
  • Utilized Python to scrape comments from YouTube and Malaysia automotive news website.
  • Orchestrated pre-processing tweets flow with noise classifier model to get clean text data.
  • Created multi-lingual sentiment analysis model using FastText to identify sentiment for each tweet.
  • Presented findings to MIROS using Streamlit and published a paper.
October 2020 - April 2021

Educations

Universiti Sains Malaysia

Master of Science
Computer Science - Intelligent System Techniques

Research Mode

January 2022 - January 2024

Universiti Teknikal Malaysia Melaka

Bachelor of Computer Science
Computer Science - Artificial Intelligence

CGPA: 3.65

September 2017 - April 2021

Skills

Programming Languages & Tools
  • Python
  • Pytorch
  • Django
  • AWS EC2
  • PostgreSQL
  • TensorFlow
  • Raspberry Pi
  • Arduino
  • Flutter

Projects

TriageAI

Implementing AI to ED Triage
Tools: PyTorch, Django

Improved an existing method described in a paper that employed BERT, TabNet, and CNN by substituting the CNN component with a Transformer Encoder. The resulting model demonstrates improved F1-score by 3%.

Social Media Sentiment Analysis

Analysis about road safety sentiment in Malaysia
Tools: Python, FastText, Streamlit

Scraped data from YouTube and collected Twitter data using Twitter API. The text data are processed and cleaned. Then, the findings were displayed using Streamlit. The preprocessing steps were published in International Journal of Road Safety.

Protecc

A wearable to care for elderly
Tools: Python, TensorFlow, Flutter, Firebase, MQTT, Raspberry Pi, ESP 32, MPU 6050

Utilized TensorFlow to create LSTM for Human Activity Recognition (HAR). Then, using TensorFlow Lite to optimize the model for Raspberry Pi. The HAR data are collected using ESP32 with MPU 6050 as sensor. Then, using MQTT, the data is transmitted from ESP 32 to Raspberry Pi. Raspberry Pi process the data with lite version of LSTM then upload the prediction to Firebase. Finally, Flutter app will fetch the new data and update the UI.

Elpida

Robots and Apps to care for elderly
Tools: TensorFlow Lite, TypeScript, Raspberry Pi, Firebase

I spearheaded a collaborative project where our team successfully developed two robots, an Android app, and a Telegram Bot. My primary responsibilities encompassed designing and implementing the Telegram Bot, as well as creating a robot dedicated to accompanying the elderly and another robot focused on medication reminders. The robot that follows the elderly around uses TensorFlow Lite to extract key points from the human body and then uses XGBoost for pose estimation. My team and I collected and processed the data ourselves because the model will be sensitive to the camera placement.


Open Source Contributions

  • Ellzaf ML - Created python package for ML implementations based on research paper that I created by myself or modified from others.
  • Huggingface - My 5 fine-tuned text classification models obtained a total of more than 600 downloads.
  • Malaysian Dataset - Scraped three websites to collect over 100MB of text data for Malay LLM.
  • Ivy - Wrapped PyTorch function inside Ivy for ML Frameworks unification.

Awards

  • Gold Award - Jejak Inovasi UTeM 2020
  • 1 st Place - Artificial Intelligence Workshop 2
  • Consolation Award - National Advance Robotic Competition(NARC) 2018
  • 2 nd Place - Autonomous FV 2018
  • Consolation Award - The CODE 2018
  • Consolation Award - FaceHack 2017

Certificates

  • [Coursera - Stanford Online] AI in Healthcare Specialization
  • [Coursera - Stanford Online] AI in Healthcare Capstone
  • [Coursera - Stanford Online] Evaluations of AI Applications in Healthcare
  • [Coursera - Stanford Online] Fundamentals of Machine Learning for Healthcare
  • [Coursera - Stanford Online] Introduction to Clinical Data
  • [Coursera - Stanford Online] Introduction to Healthcare
  • [Coursera - DeepLearning.AI] AI For Medical Treatment
  • [Coursera - DeepLearning.AI] AI for Medical Diagnosis
  • [Coursera - DeepLearning.AI] AI for Medicine Specialization
  • [Coursera - DeepLearning.AI] AI in Medical Prognosis
  • [Udemy - Dr. Angela Yu]The Complete 2020 Flutter Development Bootcamp with Dart

Activities

  • JomLaunch 2022
  • Shopee Code League 2022
  • Jejak Inovasi UTeM 2020
  • CHIPTA 2020
  • Artificial Intelligence Workshop 2
  • Projek Rekabentuk Bersepadu MTUN IDP CUP (2019)
  • RBTX 2019
  • AI Summer School Singapore 2019
  • Trainer for PHP Workshop (2019)
  • Bowling Sukan Piala Dekan (2019)
  • Judge for Roborace 2019
  • Vice Program Director for Roborace 2019 (2019)
  • AVIATAR (2019)
  • Trainer for Arduino Class (2018)
  • Melaka IOT (2018)
  • AICKATHON (2018)
  • Judge for National Roborace Challenge 2018
  • Trainer for Robotic Camp (2018)
  • President of AI Club (2018/2019)
  • NARC 2018
  • Autonomous FV 2018
  • The CODE 2018
  • BITI Workshop Series: Introduction to Git and Github
  • BITI Workshop Series: Introduction to Python
  • BITI Workshop Series: Data Analysis using RapidMiner
  • Committee member for UTeM Robocon 2017
  • Roborace 2017
  • SumoBot 2017
  • FaceHack 2017

Publications

Data Gathering and Preparation for Social Media Data Utilisation in Road and Traffic Safety

Nov 2021

Interests

Apart from being a developer, I enjoy cooking, watching TV series or movies and reading books. I usually bake cookies or cakes, but one of my craziest cook was making beef wellington. Some of the TV series that I truly enjoy are Mr Sunshine, Dark and Violet Evergarden. I also enjoy watching cooking series such as MasterChef Australia. Some of the movies that I enjoy are ConHeartist, Eternal Sunshine of The Spotless Mind and Everything Everywhere All at Once. I enjoy philosophical and self improvement read and sometimes I listed to podcast from Naval.