~4 years building large-scale AI systems. LLMs, RAG, MLOps, multilingual NLP. Government AI policy contributor — National AI Policy of Bangladesh (2026–2030).
Machine Learning Engineer with around 4 years of experience designing and deploying large-scale AI systems. Deep expertise in LLMs, Transformer architectures, RAG pipelines, and AI agents.
Strong proficiency in Python, PyTorch, HuggingFace, and MLOps. Proven track record in multilingual and low-resource NLP, model fine-tuning and optimization (LoRA, quantization), and building scalable, production-grade ML platforms.
Experienced in system-level AI strategy and cross-functional collaboration. Government AI policy contributor — helped draft Bangladesh's National AI Policy (2026–2030) and supported technical assessment of NDGA & PDPO Ordinances.
EBLICT-Dream71 · Bangladesh Computer Council, Gov't of Bangladesh
Ministry of Posts, Telecom & IT · Office of the Special Assistant, GoB
Bengali.AI · Remote
Synesis IT PLC · Dhaka, Bangladesh
Efficient grapheme parsing algorithm + normalizer handling 8 normalization issues in large Indic corpora. Published at the 2024 Joint International Conference on Computational Linguistics.
View Paper →Binary classifier hybridizing classical ML and deep learning algorithms for Bangla fake news detection.
View Paper →ML models predicting COVID-19 mortality likelihood for pregnant women using classical learning algorithms.
View Paper →System-oriented technical reviews and standards analysis. Contributed inputs ensuring feasibility, interoperability, and alignment with national digital infrastructure.
Technical assessment supporting establishment of a data governance authority. Aligned provisions with AI ecosystem engineering requirements and interoperability standards.
Technical review aligning PDPO requirements with practical AI ecosystem development while safeguarding citizen privacy rights.
Flask-based annotation platform supporting classification, NER, QA generation, and coreference resolution for NLP dataset creation.
Multi-level GAN architecture for rain streak removal from images with progressive artifact elimination and perceptual loss.
ML models using n-gram features for detecting offensive and toxic content in social media posts for content moderation.
TensorFlow.js-based real-time object detection running entirely in the browser via webcam input — zero server dependency.
NLP-based clinical question-answering chatbot leveraging HuggingFace transformer models for medical information retrieval.
Highest Distinction Graduate
BRAC University · 2024
Dean's List ×3 & VC's List ×3
BRAC University · 2020–2024
Merit Scholarship ×5
BRAC University · 2020–2024
Champion — Quarantine Contest
BRACU · 2020
Champion — Programming Bootcamp
BRACU · 2020
BRAC University, Bangladesh
Open to research collaborations, ML engineering roles, and AI governance consultations.
quaziadib@gmail.com
GitHub
github.com/quaziadib
linkedin.com/in/quaziadib
Google Scholar
Research Profile
+880170-877-0569