Journey
TUBITAK Sayem Project
Ankara
- Building an autonomous robotic system for intelligent farming, designed to operate reliably in unstructured outdoor field conditions.
- Combining perception, planning, and control modules into a unified pipeline for automated field operation.
- Automating crop monitoring and field navigation to reduce manual labor and improve farming efficiency.
Robotics Researcher
Ankara, LIRA Lab
- Focusing on methods for inference-time obstacle avoidance in diffusion-based planners, enabling safer trajectory generation without retraining the underlying model.
AI Engineer
Mozayedegar, Tabriz University Incubation Center (1 year)
- Led a team of engineers across NLP and computer vision projects, applying advanced prompt engineering techniques to improve model reliability.
- Built high-speed image processing models optimized for low-latency deployment in production pipelines.
- Developed business intelligence dashboards and incremental learning systems to continuously improve price prediction accuracy.
Research Team Lead
University of Tabriz, Applied Deep Learning Lab (1 year)
- Created and mentored 5 intern teams, guiding them from problem definition through to working prototypes.
- Launched projects in GNN-based trajectory prediction and real-time sign-language detection, spanning both research and applied deployment.
- Directed work on diffusion-based video generation and EEG-based sleep phase prediction, bridging generative modeling with biomedical signal analysis.
Computer Vision Researcher
University of Tabriz, Applied Deep Learning Lab (1.5 years)
- Developed U-Net and GAN-based models for document restoration, targeting scanned and photographed paper documents.
- Improved downstream OCR performance by removing severe crumpling artifacts from paper images prior to text recognition.
3D Design Team-lead and AI Engineer
Boston University, USA (2 years)
- Led a team generating synthetic 3D datasets using Blender to support downstream computer vision training.
- Curated and cleaned motion and video data to support AI-driven personalized workout solutions.
- Contributed to an AI application optimized for real-time performance within the Unity game engine.