I am
Leveraging machine learning to democratize medical diagnostics in underserved regions. Focused on efficient, bias-aware AI for tropical diseases and skin lesions — building scalable tools from limited data to empower clinicians worldwide.
AAAI 2025 — Led development of mLabLLM (LLaMA 3.2 fine-tuned) for differential diagnosis (Dengue, Malaria, Chikungunya). Integrated Bayesian reasoning + LoRA + pruning to reach 82.8% Top-3 accuracy for low-cost clinical support.
ScienceDirect 2025 — Built MSLD v2.0 dataset with diverse skin tones and benchmarked models (ResNet50 + transfer learning). Prototype web app for inclusive rapid screening.
arXiv 2022 — Curated the initial MSLD dataset; evaluated pre-trained models (ResNet50), reaching ~82.96% accuracy distinguishing mpox from similar diseases.
arXiv 2024 — Co-developed BMD-HS with 864 multi-label heart sound recordings for complex valvular disease detection, bridging auscultation and AI diagnostics.
Recognition for disruptive innovation — global competitions & fellowships that highlight collaborative excellence and real-world impact in healthcare and engineering.
Team FetoSynth (BUET) — Real-time acoustic fetal heart localization & monitoring system to improve prenatal care in low-resource settings.
Team DengueDrops (BUET) — Smart IV fluid calculator for dengue patients to optimize fluid therapy and hospital management.
Recognized by the Honda Foundation for excellence in biomedical engineering and sustainable innovation; grant support for international exchange and research.
Team Interplanetar (BUET) — Designed an innovative gas compression system for Mars aerial systems; 1st in Bangladesh and awarded Innovation Award (8th globally).
UIHP-funded agro-drone project (PakhiDrone BD) — precision farming & pesticide spraying adapted for Bangladesh's rural terrain; incubated into early-stage startup.
From industry distribution strategy to lab research and founding startups — practical execution and operational impact across domains.
Managing a 420mn BDT annual distribution; developed satellite & geo-data model for rural 'Char' areas and built an automated shipment app for route-to-market efficiency.
Built real-time data collector with Raspberry Pi attachments, contributed to medical LLM for differential diagnosis, and created 3D landmarks from sparse 2D video for low-compute imaging improvements.
UIHP-funded startup building agro-drones for precision farming and pesticide spraying, optimized for Bangladesh's rural terrain and smallholder needs.
Combining technical precision with creative storytelling — award-winning short films, VFX, and NFT work through a dedicated studio.
Founded studio for 3D animation, NFTs and VFX; produced award-winning short films and immersive visual narratives that blend engineering precision with artistic expression.
Pookie Men
Lectures on LLM
1st Mars Rover from BUET
Recognition for
International Award
Director of JICA
Director of Honda
Foundation
Exclusive Visit to Honda
Engineering
Exploring Culture
Study Tour in Japan
With Health Advisor of BD
Work with ImpactHub
Opening Speech in BUET
Fest
YY Ventures
Case Competition
CBID Scene
Young Engineer and
Scientist Recognition
Skin Cancer Detection
Project
Datathon on Heart Sound
RTM Tool for BAT
UIHP Pakhi Drone
DengueDrops Development
Team
IPAS Featuring_low.webp)
Recognition for Drone of
Mars (IPAS)
SciFi Writing
URC
Game based on Realtime Geo
Data
Game for IOS and Android
Platform
3D Scene x Game
3D Game Development in
Unreal Engine
Complex 3D Scene
Creation in Blender
Inverse Kinematics in
Blender