About
I am an AI researcher and backend engineer with hands-on experience across quantum machine learning, federated learning, large language models, and edge AI. My work bridges cutting-edge research and production deployment — from variational quantum circuits and post-quantum cryptography to on-device TFLite inference and LLM-powered applications.
I currently work as a Backend Engineer (AI/ML) at MAEKNIT (New York, remote), building AI-enabled backend systems and integrating LLM APIs. I am also the founder and research lead of ZogBiyog Research Lab, where I lead initiatives in Quantum ML and Post-Quantum Cryptography. Previously, I interned at PLIMES Inc. (Japan) optimising on-device models for the GOKURI edge AI pipeline.
My research focuses on trustworthy AI in adversarial and resource-constrained settings: quantum-assisted medical imaging, privacy-preserving dementia classification, phishing and smishing detection in Bengali, and LLM KV-cache security using Module-LWE. I am actively seeking PhD positions in AI, quantum machine learning, and cybersecurity.
Technical Skills
Publications
10 papers · 100 total citations · Live from Google Scholar
2024 13th International Conference on Electrical and Computer Engineering
2025 3rd International Conference on Intelligent Systems, Advanced Computing
2025 International Conference on Electrical, Computer and Communication
2025 3rd International Conference on Intelligent Systems, Advanced Computing
2025 2nd International Conference on Next-Generation Computing, IoT and
International Conference on Advanced Network Technologies and Intelligent
Mendeley Data, V2
2025 2nd International Conference on Next-Generation Computing, IoT and
International Conference on Data Science, AI and Applications, 107-117
Available at SSRN 5669250
Manuscripts Under Review
4th Int. Conf. on Computing Advancements (ICCA 2026) — under review
ZogBiyog Research Lab — under review
Experience
- Design and build scalable, high-performance backend systems, RESTful APIs, and microservices using Python (Django/FastAPI) and Node.js, backed by PostgreSQL, MySQL, and MongoDB.
- Integrate large language model APIs (OpenAI, Anthropic Claude, and other AI platforms) to deliver intelligent, AI-driven product features.
- Leverage AI-assisted development tools (GitHub Copilot, Cursor) to accelerate delivery while maintaining clean, modular, well-tested code.
- Collaborate with Product, Design, and QA across the software development lifecycle; contribute to code reviews and agile ceremonies; and resolve performance bottlenecks, security vulnerabilities, and production issues.
- Lead research initiatives in Artificial Intelligence, Quantum Computing, Quantum Machine Learning, and Post-Quantum Cryptography.
- Authored and co-authored multiple peer-reviewed publications accepted at IEEE and Springer conferences.
- Improved on-device swallow and cough detection accuracy of the INDRA model via hyperparameter optimization, audio preprocessing, and spectrogram-based data augmentation.
- Built TensorFlow Lite conversion and post-training quantization pipelines for low-latency inference on resource-constrained edge and Android devices.
- Containerized end-to-end training and evaluation with Docker and automated TFLite export within the GOKURI edge-deployment pipeline.
- Built a hybrid intrusion-detection model combining 1D-CNN feature extraction with Random Forest and MLP classifiers for multi-class DDoS detection on CIC-DDoS2019, integrated with Snort for real-time mitigation; published at ANTIC 2023 (Springer).
- Awarded the university's Outstanding Thesis / Project of the Year (2024).
- Designed, developed, and optimized high-performance websites, improving load speeds and user experience.
- Aligned web solutions with business goals to drive engagement and conversions.
- Built custom WordPress themes and plugins for enhanced functionality and SEO.
- Coordinated with content creators to deliver responsive, brand-aligned sites.
Education
Projects
AI-powered automation tool that searches and applies for jobs on LinkedIn based on a given resume.
Machine learning model to detect SMS phishing (smishing) using natural language processing and transformers.
CNN-based system for detecting plant diseases using image processing with TensorFlow and PyTorch.
Real-time FPL leaderboard that fetches live match data using the Fantasy Premier League API.
AI assistant for finance tasks with chat, voice commands, and file upload support using Google Gemini and Groq Whisper.
Awards & Recognition
Get In Touch
Open to PhD positions in AI, quantum machine learning, and cybersecurity. I also welcome research collaborations, engineering roles, and academic discussions.
gazitanbhir@proton.me