name : M. Rifqi Dzaky Azhad
degree : B.Sc. Informatics — Telkom University (GPA 4.0 / 4.0)
focus : Medical Imaging AI · Bioacoustics · Multimodal Deep Learning
currently: Fatigue detection (Trans Track) · Deepfake detection research · VLM for chest X-ray| Project | Highlight | Status |
|---|---|---|
| Bone Metastasis Diagnosis | YOLOv8 + nnU-Net v2 + Radiomics — patient AUC 0.956, beats EXINI & BONENAVI | Published |
| Bird Sound Identification | MixIT + BirdNET v2.4 — macro F1 +44% on endemic/endangered species | Presented @ ICITACEE 2025 |
| Multimodal Deepfake Detection | VideoMAE + Whisper + DANN gated MoE — intra F1 0.94, cross-dataset F1 0.60 | In progress |
| VLM Chest X-Ray Explanation | BiomedCLIP + structured radiology-style text — radiologist usability evaluation | Proposal |
Core ML
Computer Vision & Medical AI
Engineering
|
research-deepfake_classification Modality-Aware MoE deepfake detector. Cross-modal attention gate routes each clip to AudioSyncExpert or VisualArtifactExpert without knowing manipulation type at test time.
|
Research_Birdsound-Classification MixIT + fine-tuned BirdNET for 219 species in Borneo. +44% macro F1 on endemic and endangered species vs. baseline.
|
|
Async fatigue alarm review API for Trans Track MDVR fleet system. Accuracy 0.85, macro F1 0.78 at 10 FPS.
|
aplikasi_berbasis_platform-retogen Full-stack article management platform with JWT auth, nested comments, and 185 automated tests.
|