Winy Ainia

• Built industrial AI products from scratch for mining use cases, including computer vision-based safety monitoring and NLP-based emergency reporting systems. • Led end-to-end technical execution across data collection, dataset preparation, model training, inference pipeline development, deployment architecture, and implementation readiness. • Served as the primary technical owner for key AI initiatives, driving technical decisions across AI architecture, deployment strategy, integration approach, and operational readiness. • Designed and deployed AI services on NVIDIA Jetson Orin NX and AWS EC2, optimizing pipelines for real-time and resource-constrained environments. • Implemented MLOps and production ML practices, including deployment workflows, monitoring, alerting, logging to database, and CI/CD using GitHub Actions. • Built and integrated backend services using FastAPI, API-based workflows, Dockerized deployment, and cloud/edge runtime environments. • Managed full AI inference pipelines from RTSP stream ingestion → AI processing → HLS/web-accessible output for production monitoring systems. • Standardized deployment and delivery using Docker, Docker Compose, Harbor, MinIO, and Linux-based environments. • Configured and integrated industrial camera systems for field deployment. • Mentored a junior teammate and led technical execution in practice.

• Evaluated and ranked AI-generated responses to improve performance and verify factual accuracy. • Provided key feedback to refine AI-generated content, improving language fluency and contextual relevance.

• Analyzed AI model outputs by classifying daily images into multiple risk categories, ensuring model reliability and safety compliance. • Provided detailed, data-driven feedback to the machine learning team, contributing to notable model accuracy and performance enhancements through annotation insights. • Performed high-precision image annotation using bounding boxes and segmentation, consistently achieving up to 95% - 100% accuracy to improve training data quality.

• Successfully disseminated bachelor thesis research at an international seminar, receiving valuable expert feedback while demonstrating advanced problem-solving skills. • Enhanced English presentation skills by effectively conveying complex concepts simply and engagingly. • Publishing the research paper in Taylor & Francis

• Successfully coordinated a diverse team of 33 members to develop a cutting-edge underwater robot prototype within a tight 3-month timeline and limited budget, advanced to the finals. • Spearheaded the strategic planning and execution of division objectives, resulting in significantly higher operational efficiency compared to the previous period and improved team collaboration. • Successfully managed project schedules, budgets, and technical challenges, ensuring timely completion of all tasks.

• Pioneered the official establishment of ROBOTIIK as a faculty institution, developing its legal framework. • Successfully implemented a more efficient recruitment system by providing a month-long training program for potential members to identify and develop their potential. • Conducted regular performance evaluations and provided constructive feedback to all members.

• Designed and developed an Autonomous Underwater Vehicle (AUV) prototype for search and rescue operations in maritime accidents, integrating deep learning for object detection through computer vision, autonomous Pixhawk technology, and equipped with sensors and thrusters from Blue Robotics. • Reached the final round, securing 12th place out of 44 teams, demonstrating strong technical skills and innovation. • Successfully presented the robot live to the judges, showcasing practical innovation and performance.