At a Glance

Background

B.Sc. Computer Science graduate from BRAC University, Dhaka
Specialization in Machine Learning, Deep Learning, Computer Vision & Prompt Engineering.

Location & Status

Based in Dhaka, Bangladesh.
Fresh graduate, actively open to ML engineering roles, AI research positions, and collaborative projects — local and remote.

Career Goal

Seeking entry level roles in a software company where I can contribute meaningful work and continue growing as a researcher and engineer.

Contact

mohammodtareqaziz@gmail.com
write2justice@gmail.com
📞 +88 01922-667040
📞 +88 01575-616172
📍 Merul Badda, Dhaka-1212

Zero-Shot Anomaly Detection — Thesis

• Dual-stream spatiotemporal framework
• TimeSformer + CLIP + DPC-RNN
• 84.5% ROC-AUC on UCF-Crime dataset
• Published to ArXiv

Blue-Light Glasses via ML — CSE 424

• Spectrophotometric data, 50 lenses
• KNN model: 91.4% R² score
• Compared SVM, KNN, Linear Regression
• IEEE-format paper

Clustering & Stability Analysis — CSE 425

• Novel VAE-inspired SCNN architecture
• Compared K-Means, GMM, SOM
• Uncertainty quantification on Wine dataset
• Multi-seed stability analysis

Loan Eligibility Prediction — CSE 427

• 5 ML classifiers benchmarked
• RF & MLP: 89.29% accuracy
• SMOTE for class imbalance
• Full preprocessing pipeline

Zero-Shot & Self-Supervised Learning

Building models that generalize to unseen classes without explicit supervision — through contrastive and predictive learning.

Spatiotemporal Video Understanding

Architectures that model both spatial structure and temporal dynamics in video — applied to surveillance, behavior analysis, and anomaly detection.

Vision-Language Models

Multi-modal learning that aligns visual and linguistic signals for richer semantic scene understanding — as demonstrated in my use of CLIP.

Applied ML & Data-Driven Diagnostics

Using ML for real-world problems in health informatics, anomaly detection, and financial prediction — where reliability and interpretability matter.