👨🏻‍💻 Hello there👋🏼!! I’m an Applied AI Scientist with senior-level experience translating research into real-world impact across healthcare, finance, and cybersecurity. I completed my PhD in Information Systems (AI & Machine Learning) at the University of Maryland, Baltimore County in April 2026.

💡 Research Highlights My doctoral research focused on building AI systems for personalized support in online health communities:

• PL-NCF (Pseudo-Label Neural Collaborative Filtering) – Invented a novel framework addressing extreme interaction sparsity in online health recommendations, achieving a 2× improvement in HR@5 and an 83.4% relative gain in NeuMF-PL variants.
• gDMR & gSTM Topic Models – Architected probabilistic topic models integrating user-generated text, demographic metadata, and Node2Vec network embeddings, achieving 15% coherence improvement and 2× topic distinctiveness over LDA, DMR, and STM baselines across 4 datasets.
• Large-Scale Evaluation – Applied rigorous statistical validation on datasets with 2M+ users and 8M+ posts across 15 health forums (Cohen Kappa inter-rater reliability κ = 0.78).
• Open-Source Contributions – Released an R package for gSTM enabling researchers to implement these models in diverse projects.

🎓 Academic Background

• PhD, Information Systems (AI & ML) – University of Maryland, Baltimore County (2026)
• MS, Statistics – University of Kentucky (2021)
• MS & BS, Statistics – University of Dhaka (2017, 2015)

💼 Industry Experience

Applied AI Scientist – Behavioral Health | Technuf LLC (Dec 2025 – Apr 2026) Designed a PPO-based RL recommendation engine that increased session retention by 15% and user engagement by 10% across 5 behavioral health programs. Operationalized responsible AI auditing (SHAP + Fairlearn), reducing bias gaps by 22% across 3 protected demographic groups. Directed compliant data pipelines processing 1M+ records daily under HIPAA and GDPR.

AI Research Scientist, Edge Intelligence | Axiado Corporation (Jul 2025 – Nov 2025) Pioneered lightweight transformer-based temporal modeling for embedded cybersecurity, improving threat detection accuracy by 20% and reducing false positives by 12% on ARM Ethos-U NPUs. Achieved 4× model size reduction via INT8 quantization-aware training and structured pruning with <1% accuracy loss.

Applied AI Scientist – Generative AI & Threat Intelligence | Technuf LLC (Jun – Aug 2024) Designed RAG pipelines integrating 50K+ threat intelligence documents, accelerating incident response by 40% and reducing mean-time-to-resolution by 25%. Built hybrid threat-scoring models that reduced low-value alerts by 30%.