Seeking a postdoc or ML engineer role with a PhD in Medical Informatics, strong Python/SQL and problem-solving skills (686+ LeetCode), and expertise in ML-based system identification, Agentic RAG, LLM agents, and sparse tensor research (GCXS in pydata/sparse):
- PhD in Medical Informatics (Vrije Universiteit Brussel, Belgium) with a focus on fast, noise-resilient system identification techniques using machine learning, published in top venue like IEEE transactions and Elsevier Q1 journals.
- Strong problem-solving skills in Python and SQL, with 686+ LeetCode challenges solved, demonstrating expertise in algorithms, data structures, and scalable full-stack development.
- Expert in Machine Learning, Agentic RAG, and Data Science, currently building Agentic RAG pipelines and autonomous LLM agents for domain-specific document retrieval and contextual reasoning.
- Significant impact of research to Sparse Data Structures, with a Master’s thesis on high-dimensional tensor representation adopted in the pydata/sparse Python library (GCXS format), enabling efficient sparse computation.
- Scientific contribution and community involvement as a reviewer for IEEE Transactions on Instrumentation & Measurement and Mobile Computing, a TPC member for ISCC 2024–2025, and Technical Co-chair for the AI track at EICT 2023 and 2025.
Education
-
PhD in Medical Informatics (2017-2022)
Biostatistics and Medical Informatics research group
Vrije Universiteit Brussel, Belgium
Supervisor: Prof. Dr. Kurt BARBE
Thesis: Wiener-Hammerstein System Identification applied to bioimpedance spectroscopy: Computationally fast model configuration
Abstract: Thesis proposes new techniques to minimize blood loss during surgical operation. It is done by the proper mathematical formulation of the bio-impedance on the electrosurgery data from Covidien Inc. The study hypothesized the underlying behavior as Wiener-Hammerstein (W-H) system. The computation time to identify the W-H system is minimized to a few minutes from several hours without any prior knowledge on Benchmark data. The good initial models are selected based on the Spearman correlation between internal signals of W-H system. But the performance of this technique is affected by the noise level. It can be recovered by a proposed variation of popular random forest technique.
-
Masters in Computer Science (2012-2016)
Khulna University of Engineering & Technology, Bangladesh
Supervisor: Prof. Dr. K.M. Azharul Hasan
CGPA: 3.58 (out of 4.00 in Course-work)
Thesis: An Efficient Representation of Higher Dimensional Arrays and Its Evaluation
Abstract: Thesis proposes the matricization of higher dimensional data for high performance computing. It is done by fitting odd numbered dimensions along row direction and even dimensions along column direction. It saves the storage and computation of highly sparse tensors.
Achievement: A popular python library ‘pydata’ is using above thesis to compute sparse multi-dimensional array as “This is stored in GCXS format, a generalization of the GCRS/GCCS formats from 'Efficient storage scheme for n-dimensional sparse array: GCRS/GCCS'”.
-
Bachelor in Computer Science (2005-2009)
Khulna University of Engineering & Technology, Bangladesh
Thesis/Final Project: KUET CRTS System & Online Registration and Result Processing (ORRP)
Research Interest
- Medical Informatics
- High Performance Computing
- Large Language Model (LLM)
- System Identification
- Advance Control System
- Agentic RAG
Journal Publication
- M.A.H. Shaikh, and K. Barbé, "Dynamical System Modeling to Discriminate Tissue Types for Bipolar Electrosurgery," Elsivier Biomedical Signal Processing and Control, 2023. Q1 (1 citation)
- M.A.H. Shaikh, and K. Barbé, "Study of Random Forest to Identify Wiener-Hammerstein System," IEEE Transactions on Instrumentation and Measurement, vol. 70, 2021. Q1 (23 citations)
- M.A.H. Shaikh, and K. Barbé, "Wiener–Hammerstein System Identification: A Fast Approach Through Spearman Correlation," IEEE Transactions on Instrumentation and Measurement, vol. 68(5), 2019. Q1 (45 citations)
- K.M A. Hasan, and M.A.H. Shaikh, "Efficient representation of higher-dimensional arrays by dimension transformations," Springers The Journal of Super Computing, vol. 73(6), 2017. Q2 (12 citations)
Conference Publication
- M.A.H. Shaikh, and K. Barbé, "Initial Estimation of Wiener-Hammerstein System with Random Forest," Proc. of IEEE International Instrumentation and Measurement Technology Conference (I2MTC), Auckland, New Zealand, 2019. (6 citations)
- M.A.H. Shaikh, and K. Barbé, "Spearman correlation for initial estimation of Wiener-Hammerstein system," Proc. of IEEE International Instrumentation and Measurement Technology Conference (I2MTC), Houston, Texas, 2018. (6 citations)
- M.A.H. Shaikh, G.G.M.N. Ali, P.H.J. Chong, and Y.L. Guan, "Parallel matricization for n-D array operations," Proc. of IEEE Region 10 Conference (TENCON), Singapore, 2016.
- M.A.H. Shaikh, K.M.A. Hasan, G.G.M.N. Ali, M. Chafii, and P.H.J. Chong, "Efficient Matricization of n-D Array with CUDA and Its Evaluation," Proc. of IEEE Intl Conference on Computational Science and Engineering (CSE), Paris, France, 2016. (3 citations)
- M.A.H. Shaikh, M.T. Omar, and K.M.A. Hasan, "Efficient index computation for array based structured data," Proc. of International Conference on Electrical Information and Communication Technologies (EICT), Khulna, Bangladesh, 2015. (4 citations)
- M.A.H. Shaikh, and K.M.A. Hasan, "Efficient storage scheme for n-dimensional sparse array: GCRS/GCCS," Proc. of International Conference on High Performance Computing & Simulation (HPCS), Amsterdam, Netherlands, 2015. (17 citations)
- K.M.A. Hasan, and M.A.H. Shaikh "Representing Higher Dimensional Arrays into Generalized Two-Dimensional Array: G2A," Proc. of Advances in Parallel and Distributed Computing and Ubiquitous Services (PDCAT), Jeju, South Korea, 2015. (10 citations)
Scholarship and Award
- Doctoral Fellowship, Vrije Universiteit Brussel, Belgium (2017-2020)
- Student Travel Grant, HPCS 2015, Amsterdam, The Netherlands (July 2015)
- Technical Scholarship, Khulna Univ. of Eng. & Technology, Bangladesh (2007-2008)
- BSC Scholarship, Nippon Foundation, Japan (2006)
Reviewer
- IEEE Transactions on Instrumentation and Measurement (Since 2019)
- IEEE Transactions on Mobile Computing (Since 2022)
- IEEE TENCON, Singapore, 2016
- IEEE ISCC, France, 2024
TPC Member
- IEEE ISCC, France, 2024
- IEEE ISCC, Bologna, 2025
Technical Co-Chair
- International Conference on Electrical Information and Communication Technology (EICT 2023), Track- Artificial Intelligence
- International Conference on Electrical Information and Communication Technology (EICT 2025), Track- Artificial Intelligence
Job Experience
- System Analyst at Khulna University of Engineering & Technology (Feb 2022-Current):
- Designed and implemented an Routine Planner for examination scheduling using genetic programming, reducing manual effort by 90%. It can schedule Examination in minimal days with good chromosome.
- Conducted training sessions on Data Science and Machine Learning with Python for 100+ diploma students, enhancing their technical proficiency.
- Database Programmer at Khulna University of Engineering & Technology (Sep 2016-Jan 2022):
- Task 1
- Task 2
- Assistant Programmer at Khulna University of Engineering & Technology (Jun 2012-Aug 2016):
- Task 1
- Task 2
- Assistant Programmer at HEQEP Project of KUET (April 2011-May 2012):
- Task 1
- Task 2
Technical Experience
Solved 686+ algorithmic problems on LeetCode, showcasing exceptional problem-solving and coding proficiency:- Machine Learning: Agentic RAG, LLM, Random Forest, SVM, Neural Networks, Feature Engineering, Model Optimization
- Data Science: Pandas, NumPy, Scikit-learn, Matplotlib, PowerBI, Tableau
- Programming: Python, C/C++, SQL, PHP
- Databases: PostgreSQL, MySQL (Database Design, Optimization, High-Performance Querying)
- Backend Development: FastAPI, Flask, REST APIs, Microservices Architecture
- Frontend Development: React.js, Alpine.js, HTML/CSS, JavaScript
- DevOps & Deployment: Docker, CI/CD Pipelines, Linux (Ubuntu, Debian)
Current Project
- Text Descriptor (ongoing)
The project aims to predict the context from cartoon. We are training cartoon dataset from Google with LLaMA model.
- Agentic RAG based Institutional Chat
The KUET Chat System is an AI-based information provider, developed by me using Retrieval-Augmented Generation (RAG) and Agentic AI. The Agentic AI feature makes the bot act like a guide, helping students and teacher complete tasks like course registration or finding informations automatically. It helps users get instant answers and automate common tasks of the university. The core of the system uses RAG to find and combine information from documents, giving accurate answers based on context. The backend uses FastAPI for handling many requests at once, and the frontend ensures a smooth user experience with Next.js and React:
- Front-End: Next.js
- Back-End: FastAPI
- LLM Model: llava-1.5-7b
- Vector DB: Chromadb
- Gen-AI: Py-torch
Completed Project
The project aims to predict the context from cartoon. We are training cartoon dataset from Google with LLaMA model.
The KUET Chat System is an AI-based information provider, developed by me using Retrieval-Augmented Generation (RAG) and Agentic AI. The Agentic AI feature makes the bot act like a guide, helping students and teacher complete tasks like course registration or finding informations automatically. It helps users get instant answers and automate common tasks of the university. The core of the system uses RAG to find and combine information from documents, giving accurate answers based on context. The backend uses FastAPI for handling many requests at once, and the frontend ensures a smooth user experience with Next.js and React:
- Front-End: Next.js
- Back-End: FastAPI
- LLM Model: llava-1.5-7b
- Vector DB: Chromadb
- Gen-AI: Py-torch
