I am a machine learning engineer and applied AI researcher with hands-on experience building end-to-end systems
across computer vision, deep learning, and geospatial intelligence.
My work spans a range of real-world projects, including virtual try-on systems for fashion and cosmetics,
AI-driven ERP assistants, and urban land use classification using satellite imagery. I specialize in deep
learning, MLOps, and human-centric applications of computer vision. With a strong foundation in both research and
engineering, I focus on creating scalable, impactful AI solutions grounded in thoughtful design and technical
rigor.
Research interests: Machine Learning, Deep Learning, Computer Vision, GIS, and Remote Sensing
Get in touch
Feel free to reach out at anissarker603@gmail.com for
collaborations, research discussions, or
just to connect.
Contributed to the development of an AI assistant for ERP systems by designing a natural language interface
for querying business data. Applied LLMs and internal orchestration frameworks to bridge user intent with ERP
logic. Created custom data visualizations to improve data comprehension for stakeholders, boosting reporting
efficiency.
DropDead Gorgeous: Virtual Makeup Room
Led the development of a real-time virtual makeup application using facial landmark detection and computer
vision. Focused on applying dynamic effects like lipstick and blush with high visual fidelity, ensuring a
fluid and interactive user experience.
Virtual Try-On for Clothes
Engineered a virtual try-on system enabling users to visualize clothing on their body in real-time. Utilized
pose estimation and body segmentation to align garments naturally with user movements and body shapes. Focused
on ensuring accurate garment overlay, smooth transitions, and visual realism across diverse poses. Improved
user engagement and interactivity through responsive and lightweight implementation.
Urban Categorization with Deep Learning
Designed and trained semantic segmentation models for satellite imagery to classify urban land use in Dhaka.
Managed data pipeline preparation, annotation, and preprocessing. Collaborated with international partners to
ensure geographic and model generalizability.
