Machine Learning Engineer

Skills

• Programming Language: Python, C/C++, Java, Latex
• ML Framework: PyTorch, PyTorch Lightning, Tensorflow, Keras, Huggingface
• Library: Numpy, SciPy, Pandas, SkLearn, Librosa, OpenCV, NLTK

Education

• Bachelor of Science in Computer Science and Engineering
• Chittagong University of Engineering and Technology (March 2022)

Knowledge Set

• Core Skills
  • Signal Processing
  • Numerical Analysis
  • Probability and Statistics
  • Machine Learning
• Applied Domains
  • Natural Language Processing
  • Speech Signal Processing
  • Multi-Modal Representation Learning
  • Self-Supervised Learning
  • Data Compression
• ML Architectures
  • Variants of Transformers (BERT, GPT, Conformer, etc)
  • Generative Modeling (Normalizing Flows, VAE, GAN, WGAN)
  • Neural ODE
  • RCNN Family
  • UNET, CNN, LSTM

Research Interest

Multimodal Representation Learning, Embodied AI, VLM, Physics Informed Neural Network

Work Experience

NLP Engineer(Speech and Natural Language) @ Hishab Technologies Limited (September 2023 - Present)

• Conversation AI
• Few-show Text to Speech Synthesis using GPT2 and Neural Audio Codec
• Normalizing Flow-based Realtime Bengali Speech Synthesis
• Multilingual NMT model to translate Indic to English Language and vice-versa

AI Engineer @ Next Soluion Lab (July 2022 - July 2023)

• Hierarchical Graphical Relation Extraction from Unstructured Document: A Transformer-based Multimodal-Multitasker Model capable of capturing multi-level spatially-related semantic relationships among non-homogeneous entities from the predicted entity relationship(adjacency) matrix.

Associate AI Engineer @ Next Soluion Lab (July 2021 - July 2022)

• Table Structure ReGeneration for Structured Document: Adapting Wireframe Parsing framework in regenerating irregular tabular object for an OCR system.
• Text Orientation Correction: Utilizing Data Compression along with Residual Network to correct document orientation to increase the effectiveness of the OCR.

Publications

1. M. J. I. Basher, M. M. Hoque, S. Afroze.(2023). BnGraphemizer: A Grapheme-based Tokenizer for Bengali Handwritten Text Recognition.Published in 9th IEEE International Women in Engineering (WIE) Conference on Electrical and Computer Engineering 2023 (IEEE WIECON-ECE 2023.
2. M. Hasan, M. J. I. Basher, and *M. T. R. Shawon(2023). Bengali Intent Classification with Generative Adversarial BERT. Published in International Conference on Computer and Information Technology (ICCIT), 2023.

Under Review

1. M. J. I. Basher, M. R. Noor and S. Shatabda (2023). Transformer-Based Bengali Handwritten Text Recognition with Self-Supervised Contrastive Pretraining. Submitted in Computers and Electrical Engineering.
2. Md. Zubair, Md. S. Salim, M. M. Rahman, M. J. I. Basher, S. Imran, and I. H. Sarker(2023). Agricultural Recommendation System based on Multivariate Weather Forecasting Model. Submitted in The International Journal of Intelligent Real-Time Automation.

Ongoing Research

• Enhancing Color Image Security: A Fusion of Zero-Distortion Steganography and Cryptographic Algorithm
• Parallel Linguistic Representation Learning