Ph.D. Student — University of Alabama

Wasiul Haque

Software Engineer · Researcher · Tuscaloosa, Alabama

Wasiul Haque

A software engineer and researcher currently pursuing a Ph.D. in Civil Engineering at the University of Alabama, where I work at the Connected and Automated Mobility (CAM) Lab under the supervision of Dr. Mizanur Rahman. My work sits at the intersection of machine learning, language models, and autonomous systems. Prior to my doctorate, I spent over a year as a software engineer at REVE Systems building state-of-the-art Bengali speech AI pipelines, and co-founded an ed-tech startup.

mhaque16@crimson.ua.edu +1-205-246-1247 wasiul-haque wasiulhaque

Focus Areas

Research Interests

Machine Learning Large Language Models Autonomous Vehicles Computer Security Natural Language Processing Computer Vision Indoor Navigation Data Science Cybersecurity Fundamental Systems

Academic Background

Education

University of Alabama Current
Doctor of Philosophy, Civil Engineering
Jan 2024 – Present
Tuscaloosa, AL
  • GPA: 4.00 / 4.00 Perfect
  • Research: Autonomous Vehicle, Language Models, Cybersecurity, Indoor Navigation
  • Lab: Connected and Automated Mobility (CAM) Lab
Islamic University of Technology
Bachelor of Science, Software Engineering
Jan 2019 – May 2023
Gazipur, Bangladesh
  • GPA: 3.81 / 4.00 — Class Rank: 10th of 47
  • Research: Anomaly Detection, Computer Vision
  • Thesis: Joint Position-based Anomaly Detection using Graph Convolution Network

Academic Work

Research Experience

Robust Bengali Speech-to-Text & Text-to-Speech Pipeline
Industry Research · REVE Systems · EBLICT
  • Led software engineering for a crowd-sourced & scraped Bengali speech dataset pipeline.
  • Fine-tuned the multilingual Wav2Vec2 network to build a state-of-the-art Bengali STT model; built a public-facing validation and serving system.
  • Fine-tuned the end-to-end speech synthesis model VITS for Bengali TTS; deployed with public validation service.
  • Augmented training with noisy audio to improve real-world robustness.
  • Created an Analytics Pipeline with multiple metric models to evaluate generated audio quality.
Joint Position-based Anomaly Detection using Graph Convolution Network
Academic Thesis · Islamic University of Technology · CVLab
  • Addressed the resource-intensive nature of CNN-based video anomaly detection.
  • Proposed DD-Net (Double-feature Double-motion Network), a lightweight and efficient architecture.
  • Demonstrated faster anomaly prediction without compromising accuracy in video sequences.

Professional History

Experience

University of Alabama
Graduate Research Assistant · CAM Lab
Jan 2025 – Present
Tuscaloosa, AL
  • Conducting research at the Connected and Automated Mobility (CAM) Lab.
REVE Systems
Software Engineer (L2)
Jun 2023 – Dec 2024
Dhaka, Bangladesh
  • Engineered and trained state-of-the-art Bengali STT and TTS models with backend microservices.
  • Built and deployed a Bengali real-time Voice Bot converting voice input to automated speech output.
  • Developed 30+ client tools for serving Bengali STT/TTS models in JavaScript and C#.
  • Deployed production-grade end-to-end ML inference pipelines using PyTorch TorchServe.
MasterCourse
Co-Founder & Chief Technology Officer
Mar 2021 – Present
Dhaka, Bangladesh
  • Co-founded an ed-tech startup with a vision to build a comprehensive learning platform.
StreamsTech Ltd.
Software Engineer Intern
May 2022 – Sep 2022
Dhaka, Bangladesh
  • Engineered and deployed an employee management system for Bangladesh Railway.
  • Built a geographical management solution for Bangladesh Railway Track and Bridge Management Project.

Selected Work

Projects

Backend System for STT & TTS Services
Industry
  • Scalable backend with Python FastAPI and Redis, designed as independent microservices.
  • Inter-process messaging via message brokers; each service independently scalable.
Real-time Voice-Bot Backend
Industry
  • Telephone-integrated chatbot backend processing 20 ms audio chunks in real-time.
  • Converts voice commands to prompts and delivers fast automated speech responses.
ML Inference Deployment Pipeline
Industry
  • End-to-end PyTorch TorchServe inference pipeline containerized with Docker.
  • Converted checkpoints to ONNX format for faster inference via ONNX Runtime.
Ecfor — Sign Language ↔ Speech
Academic
  • Real-time ML app translating American Sign Language video into speech.
  • Bidirectional: also converts English speech into ASL via prerecorded video.

Technical Proficiency

Skills

Languages
Python C# JavaScript Java Swift
Frameworks
FastAPI ASP.Net ReactJS PyTorch
Tools & Infra
Docker Redis RabbitMQ TorchServe ONNX Runtime Cloud Computing
Research
NLP Computer Vision Wav2Vec2 VITS Graph Neural Networks
Industry
Backend Development ML Engineering DevOps MLOps

Recognition

Awards

Talent Pool Scholarship — Higher Secondary Examination
Board of Intermediate and Secondary Education, Dhaka · January 2019
Ranked 28th among 100,000+ examinees. Merit-based scholarship awarded for outstanding achievement in the Higher Secondary School Certificate Examination at the national level.

Beyond Research

Interests

🎮 Video Games 🎬 Movies 🎵 Music 💻 Coding