At Neoxi, I've been pivotal in driving a computer vision project aimed at bolstering the security of various facilities. My contributions spanned from creating innovative multi-camera object tracking and tiny object detection functionalities to maintaining a bird's-eye view of the project's needs.I orchestrated the development of an extensive MLOps pipeline, thereby streamlining data labeling, ongoing training, and assessment of the object tracker and detector models. This rigorous approach facilitated the continuous improvement of our models, ensuring they stayed relevant and consistently delivered peak performance.Taking the helm during deployment, I oversaw the integration of our final models into the customer's existing infrastructure. I meticulously ensured seamless deployment on each client's device while prioritizing compatibility and efficient execution. My holistic understanding of project needs and technical prowess led to effective solutions that not only met but often exceeded project requirements.

I joined UVEA as a remote Machine Learning Engineer and worked on Computer Vision applications for edge devices like Edge-TPU and Nvidia Jetson Family. After a few months, I was promoted to Computer Vision Team Lead. In our team, we worked on diverse computer vision projects, including:• Social Distancing Detection: Developed an AI-assisted application to detect social distancing violations in real-time using CCTV cameras. We utilized TensorFlow's SSD MobileNet V2 object detector, calculated distances between people, and deployed the model on Edge-TPU and Jetson devices. This AI engine was embedded in a SaaS product using Docker and Kubernetes. This product provides a panel to configure the cameras and models and render output video footage.• Face Mask Detection: UVEA's Face Mask Detector applies state-of-the-art computer vision algorithms to detect if people are wearing a face mask or not. To reach an accurate model that can work with cheap and low-quality cameras, we gathered a comprehensive dataset. Then tried multiple approaches to detect the faces in a scene and applied many experiments and training in TensorFlow that led us to a model with high accuracy. The final engine has been containerized using Docker.• Real-Time Pose Estimation on Edge: We optimized a PyTorch-based pose estimation model for deployment on Jetson edge devices using TensorRT. This optimized model outperforms existing solutions when applied to real-world CCTV data. Additionally, we designed an Edge-friendly pose estimation architecture in TensorFlow and trained it using the COCO dataset, further enhancing its performance and versatility.• Label-Free Object Detection Engine: UVEA offers the Label-Free Edge vision API, a service that allows for the training of specialized models tailored to specific environments. This SaaS product was deployed on AWS, utilizing Docker and Docker Compose to ensure seamless delivery of the AI engine.

Fanavard is an HR and recruitment company focused on connecting job seekers with suitable job positions. With a team comprising highly talented employees, the company aimed to empower its workforce by incorporating AI and ML concepts and tools into their skill set.During my time at Fanavard, I held the role of AI mentor for a group of young, gifted employees. Together, we tackled several company challenges, one of which involved developing a job position recommendation system for job seekers. Leveraging Keras and NLTK, we successfully designed and implemented an effective solution that provided personalized job recommendations based on individual preferences and qualifications.Through this project and my mentorship, I helped foster a deeper understanding of AI and ML among the employees, equipping them with the knowledge and expertise to tackle complex business problems. By integrating AI and ML into Fanavard's operations, we enhanced their recruitment processes and optimized the job matching experience for job seekers, ultimately leading to greater efficiency and satisfaction for all parties involved.

At IREEN, I took the lead in developing an innovative automated real estate price estimation service using state-of-the-art deep learning techniques. As the principal architect of the project's AI engine, I had a varied range of responsibilities. This began with data collection from diverse sources like Google Maps and Google Street View and transitioned into designing and implementing robust models utilizing TensorFlow.Throughout this process, an iterative cycle of experimentation and refinement allowed us to fine-tune these models for superior accuracy and performance. Our AI-driven service consistently outperformed conventional methods and produced remarkably precise estimations.One of our team's proudest accomplishments was the successful deployment of the final model on the Google Cloud platform. This milestone signaled a notable achievement for our nascent startup venture, IREEN. Presently, the project is in the fundraising phase, and we are continually working towards its advancement. Given its promising trajectory, IREEN is well-positioned for future growth and success.

Bamboo is a data analytics platform for marketing companies to monitor and manage their marketers performance.

At ISENSE, a prominent structure's health monitoring company, my primary responsibility was to create state-of-the-art data-driven models for diverse structures, such as bridges and gas pipelines. By leveraging advanced methodologies like subspace identification and Transfer Function identification methods, I successfully predicted their behavior in response to seismic activities, bolstering the overall understanding of their performance and ensuring enhanced safety measures.In addition to my technical role, I also took charge of training employees who lacked a background in data science. I effectively conveyed complex data science concepts and methodologies to enable them to contribute confidently to our projects. Through my guidance, four individuals acquired valuable skills that empowered them to excel in their roles within the company.

I was teaching assistant in various courses each semester such as Engineering Mathematics,Signals and Systems,Communication systems.