About Me
I am a Machine Learning Engineer with a strong background in NLP, data-driven modeling, and software development. I enjoy building intelligent systems that solve real-world problems, with research interests in Machine Learning, Natural Language Processing, and Large Language Models.
Work Experience
Machine Learning Engineer — REVE Systems
March 2024 – June 2025
- Developed Bengali spelling and grammar checker using pseudo-labeling and synthetic data generation.
- Created a punctuation correction model to enhance text readability.
- Applied model compression (quantization, knowledge distillation) to reduce LLM size while maintaining accuracy.
Junior Data Engineer — Wunderman Thompson
Oct 2023 – Mar 2024
- Maintained ELT processes for data movement and transformation.
- Implemented ML algorithm to fill missing values and accelerate delivery.
- Processed and transformed raw data for modeling in Snowflake.
Software Developer (Intern) — Technohaven
Jul 2023 – Sep 2023
- Built customer clustering ML model for improved loan repayment in digital banking.
- Developed secure session management with JWT and WebSocket.
- Created chatbot solution with Dialogflow CX for better customer service.
Analytics Engineer Intern — Intelligent Machines
Apr 2022 – Aug 2022
- Optimized task assignment using ML in the Bkash Biponon Project, reducing costs.
- Formulated and solved TSP for route optimization.
- Automated bank statement analysis for IDLC, cutting processing time from 3 days to 30 minutes.
Education
M.S. in Computer Science, North Dakota State University (2025 – 2027)
B.Sc. in Software Engineering, Islamic University of Technology (2019 – 2023)
CGPA: 3.71 / 4.0
Projects
E-Commerce (MERN Stack)
Web app with admin controls, product catalog, add-to-cart, and checkout functionality.
View on GitHub
Ecfor (Sign Language Translator)
Translated sign language to English using camera feed for accessibility. Built with Python, Django, and ML object detection models.
View on GitHub
Research
Bengali Spell and Grammar Checker
- Worked with mBERT, XLM-R, and translation models to design pseudo-labeling for NER tasks.
- Fine-tuned models for punctuation and spacing corrections in Bengali text.
- Injected grammatical errors to strengthen model robustness.
Technical Skills
- Languages: Python, JavaScript, C
- Databases: MySQL, MongoDB
- Machine Learning: Numpy, PyTorch, Transformers, Scikit-learn
- Frontend: React, HTML, CSS
- Backend: NodeJS, ExpressJS
Problem Solving
- LeetCode: Top 4% (Contest rating 1926), 1200+ problems solved. Profile Link: leetcode.com/afzalsiddique
- Codeforces: 165 problems solved
Certifications
Deep Learning Specialization — Coursera