(move your cursor over to interact)
About
I'm a machine learning professional from Bangladesh with a background in software engineering and 8+ years of experience, including 5+ years managing engineering teams.
Currently, I lead the Machine Learning Engineering team at Optimizely and serve as a part-time faculty member in the Department of Computer Science and Engineering at United International University (UIU) in Dhaka. Previously, I worked as a Data Science Manager at Robi Axiata Ltd and as a Data Scientist at Axiata Group Bhd.
I earned my Bachelor's and Master's degrees from BRAC University. I specialize in productizing, developing, and deploying large-scale machine learning models on big data platforms such as Cloudera, in AWS cloud environments, and on local infrastructure. I enjoy experimenting with algorithms; some of my early experiments are available at www.niloy.red. My primary programming languages are Python and Java, and I also work with Android, Spring, and Django.
Research interests
Education
BRAC University, Bangladesh
Thesis: Learning to Win: A Process Reward Model for Competitive Machine Learning Agents. Advisor: Dr. Golam Rabiul Alam
Publication · Benchmarking and Selecting Optimal Diabetic Retinopathy Detecting Machine Learning Model using Entropy and TOPSIS Method
BRAC University, Bangladesh
Thesis: Pattern Recognition with Quantum Support Vector Machine on Near-Term Quantum Processors. Advisor: Dr. Mahbubul Majumdar
Publication · Classification of Categorical Objects in Ventral Temporal Cortex using fMRI Data
Extracurricular
Professional Experience
Combining leadership in production AI/ML engineering with a commitment to university instruction, research, and community mentorship.
Europe (remote) · Full-Time
Lead a team of ML engineers and analysts developing machine learning solutions for the business.
Turing
USA (remote) · Contractual
Built LLM and RLHF/RLAIF training pipelines and Kaggle-style competition environments for Meta, targeting gold-level data science competition performance.
Dhaka, Bangladesh · Full-Time
Led teams of data scientists delivering AI projects across sales, marketing, and network operations.
Achievements · Best Innovator Award · Star Performer · Best Team Award
RedDot Digital Limited · Contractual
Dhaka, Bangladesh
Developed and deployed an end-to-end AI solution for tobacco leaf grading for British American Tobacco (BAT) by leading a team of data scientists and SDEs.
Malaysia (remote) · Full-Time
Developed machine learning models for Axiata Group operating companies, including Ncell Axiata and Smart Axiata.
Dhaka, Bangladesh · Full-Time
Developed and maintained the application backend.
Dept. of CSE, BRAC University
Dhaka, Bangladesh · Full-Time
Instructed six sections with 240+ students in Spring 2019 and four sections with 170+ students in Summer 2019.
Dhaka, Bangladesh
Raised funds from hackathons and local projects. Designed and developed 4+ e-commerce Android and web apps for SME businesses.
United International University (UIU)
Dhaka, Bangladesh · Part-Time
Designed and developed course materials for the Artificial Intelligence Lab (CSE3812) and Software Engineering Lab (CSE3422) courses.
BRAC University
Dhaka, Bangladesh · Part-Time
Co-supervised Bangla text-to-speech AI thesis work with Dr. Zavid Parvez and mentored the team through research pipelines to a peer-reviewed ACR 2023 publication.
UpSkill Ltd
Dhaka, Bangladesh · Part-Time
Built data science screening assessments for Intelligent Machines and mentored 30 software engineers pursuing roles in the AI/ML industry.
IEEE BRAC University Student Branch
Dhaka, Bangladesh · Part-Time
Built the official IEEE BRAC University web portal and led successful petitions to launch the Computer Society and Power & Energy Society chapters.
Microsoft Bangladesh
Dhaka, Bangladesh · Part-Time
Selected as the sole MSP from 200+ candidates and organized campus workshops and events on cloud infrastructure and Microsoft technologies.
Awards
Publications
In Proceedings of the 2023 International Conference on Advances in Computing Research (ACR'23)
This paper presents an approach to Bengali speech-to-text conversion using a novel dataset of 56 unique words collected from 160 participants. The LSTM model achieved 90% accuracy on an unseen dataset and was evaluated across demographic groups and noise conditions, supporting progress toward real-time Bengali speech recognition.
ICECCME 2021 · IEEE Xplore
After preprocessing a diabetic retinopathy dataset, Inception-v3 was used to extract features. Eight models were then trained and evaluated across ten metrics. Entropy-weighted TOPSIS benchmarked the models and identified AdaBoost as the best detector for the dataset studied.
TENCON 2019 · IEEE Xplore
Machine learning models applied to fMRI data from the ventral temporal cortex classified objects viewed by participants, such as cats and chairs. After hyperparameter tuning, SVM (96.92%) and k-NN (96.90%) performed best, with potential applications in brain-computer interface solutions.
BRAC University thesis · DSpace
This thesis maps classical feature points onto a quantum computer and proposes a QSVM model for near-term superconducting quantum processors. The results show that quantum optimization can train a discriminative SVM for pattern recognition.
Judging & Advisory Roles
Robi Datathon 2.0 (National Flagship Hackathon)
National
Designed technical problem statements and evaluated ML performance and code architecture across 1,685 submissions from 384 teams and 1,117 participants.
Department of CSE, BRAC University
Dhaka, Bangladesh
Adjudicated 14 undergraduate thesis defenses as an industry AI/ML expert, assessing methodology, novelty, and presentation rigor.
Hack NSU (Season 5), North South University
Dhaka, Bangladesh
Judged 11 finalist teams addressing six real-world problem statements, evaluating scalability, technical design, and execution quality.
Innovation Challenge (Season 15), North South University
Dhaka, Bangladesh
Evaluated year-long CSE, EEE, and ETE capstone projects focused on AI and machine learning to select three major award recipients.
Speaking & Instruction
Works
Selected personal projects, algorithmic experiments, and interactive pieces are available on my creative site.
Visit niloy.red to explore projects, demos, and experiments beyond this portfolio.
Visit niloy.redSkills
Programming, ML frameworks, big data platforms, visualization, and application stacks.