James Sarmiento Email & Phone Number

Singapore

○ Object Detection Expertise: Implemented and fine-tuned state-of-the-art object detection algorithms such as YOLO, EfficientDet, and Faster-RCNN. Achieved unprecedented levels of accuracy in detecting a variety of dental issues including cavities and periapical abscesses in dental x-rays.○ Hands-On Segment Anything: Created a pipeline to finetune MobileSAM using bounding boxes for a medical image segmentation domain to create an instant annotation tool model.○ Semantic Segmentation and… Show more ○ Object Detection Expertise: Implemented and fine-tuned state-of-the-art object detection algorithms such as YOLO, EfficientDet, and Faster-RCNN. Achieved unprecedented levels of accuracy in detecting a variety of dental issues including cavities and periapical abscesses in dental x-rays.○ Hands-On Segment Anything: Created a pipeline to finetune MobileSAM using bounding boxes for a medical image segmentation domain to create an instant annotation tool model.○ Semantic Segmentation and Instance Segmentation: Engineered a comprehensive training pipeline for instance and semantic segmentation models. Utilized frameworks like U-Net, DDRNet, and DeepLabV3 to successfully train models on limited datasets, surpassing industry benchmarks.○ Cloud Deployment: Seamlessly deployed trained models into Google Cloud Platform (GCP) using BentoML, facilitating real-time analysis and scalability.○ AI Research Enhancement: Innovated research methodologies by incorporating visualization techniques for dental x-rays and applying self-supervised pretraining. Achieved a 10% improvement in key performance metrics through the strategic use of unlabeled datasets.○ Regulatory Compliance & Documentation: Spearheaded the documentation process for solution designs, clinical trials, and FDA clearance, ensuring that our AI solutions meet the highest industry and legal standards. Show less

United States

Generative AI in Cybersecurity and Bot MitigationStable Diffusion, DALLE, LoRA

United States

○ Team Leadership and Mentoring: Managed and led a high-performing team of AI engineers. Provided mentorship aimed at enhancing skill sets and aligning project objectives, thereby fostering a culture of continuous innovation.○ Cross-Functional Coordination: Acted as the primary liaison between the AI R&D team and web development units. Orchestrated the seamless synchronization of API endpoints across different departments, ensuring coherent and efficient workflows.○ Strategic… Show more ○ Team Leadership and Mentoring: Managed and led a high-performing team of AI engineers. Provided mentorship aimed at enhancing skill sets and aligning project objectives, thereby fostering a culture of continuous innovation.○ Cross-Functional Coordination: Acted as the primary liaison between the AI R&D team and web development units. Orchestrated the seamless synchronization of API endpoints across different departments, ensuring coherent and efficient workflows.○ Strategic Decision-Making: Assumed the role of Acting CTO, overseeing technical strategies and decisions across the organization. Provided guidance on technology stacks, scalability, and innovation pathways, ensuring alignment with the organization's objectives.○ Image Generation Specialization: Developed and implemented a state-of-the-art image generation model service utilizing Stable Diffusion models, GANs, Image2Image, Dreambooth, Automatic1111. Utilized cutting-edge technologies such as LoRA and textual inversions for superior model performance.○ Operational Excellence: Led the deployment of machine learning models into production environments. Utilized MLOps services like BentoML for robust and efficient deployment, with a particular focus on GPU-intensive tasks. Show less

United States

○ Innovative Research: Spearheaded groundbreaking research in diverse Machine Learning sectors. Areas of expertise include Generative Models (GANs, Controllable Generation), Object Detection, Image Processing, as well as CLIP and Diffusion Models.○ End-to-End Deep Learning Pipelines: Conceived and executed comprehensive deep learning pipelines. Developed efficient processes for image clustering, filtering, and a streamlined model training workflow that resulted in high-quality models.○… Show more ○ Innovative Research: Spearheaded groundbreaking research in diverse Machine Learning sectors. Areas of expertise include Generative Models (GANs, Controllable Generation), Object Detection, Image Processing, as well as CLIP and Diffusion Models.○ End-to-End Deep Learning Pipelines: Conceived and executed comprehensive deep learning pipelines. Developed efficient processes for image clustering, filtering, and a streamlined model training workflow that resulted in high-quality models.○ Library Optimization: Re-engineered popular libraries like StyleGAN, StyleCLIP, and Stable Diffusion to achieve specific goals. Reduced manual captcha challenge generation time from two weeks to a single day through automated, infinitely variable generated challenges.○ Scalable ML Pipeline Construction: Utilized Kubernetes and Argo Workflows to architect robust, scalable machine learning pipelines. Enabled distributed text-to-image generation via diffusion models, which expedited the process of ideation to production integration from three months to one day.Expert in: Pytorch, Kubernetes, Argo Workflows, and various deep learning/computer vision libraries." Show less

Project Related to Automated Detection of Pavement Faults and Distresses on Philippine Roads with Deep Learning and Computer VisionSystem Design & Implementation: Conceived and deployed a sophisticated object detection, localization, and tracking system for Project PAVE—a 2-year initiative funded by the Department of Science and Technology (DOST). The project aims to revolutionize road infrastructure maintenance in the Philippines through the development of automated road condition survey… Show more Project Related to Automated Detection of Pavement Faults and Distresses on Philippine Roads with Deep Learning and Computer VisionSystem Design & Implementation: Conceived and deployed a sophisticated object detection, localization, and tracking system for Project PAVE—a 2-year initiative funded by the Department of Science and Technology (DOST). The project aims to revolutionize road infrastructure maintenance in the Philippines through the development of automated road condition survey equipment.Deep Learning & Computer Vision: Utilized cutting-edge Deep Learning and Computer Vision techniques to accurately identify, locate, and track various forms of pavement faults and distresses on Philippine roads.Infrastructure Impact: Contributed to the long-term objective of enhancing road safety and infrastructure durability by providing a machine-driven, data-rich approach to road maintenance and preservation. Show less

I help students and tutees with their questions about Artificial Intelligence, Machine Learning, Python Programming, Computer Science, MySQL, and Object-Oriented Programming by giving them detailed, easily understandable, easy-to-follow answers to their questions.

Philippines

○ Rebuilt entire ETL data engineering pipeline that processes hundred millions of import data from raw files to Data Lake into Elasticsearch using AWS Glue and Apache Spark○ Created an entity deduplication algorithm for matching conceptually similar company names○ Created an end-to-end text classifier for identifying persons from the list of import consignees using Spark MLLib○ Creates visualizations, finds patterns, generate insights, and formulate predictive analysis from trade… Show more ○ Rebuilt entire ETL data engineering pipeline that processes hundred millions of import data from raw files to Data Lake into Elasticsearch using AWS Glue and Apache Spark○ Created an entity deduplication algorithm for matching conceptually similar company names○ Created an end-to-end text classifier for identifying persons from the list of import consignees using Spark MLLib○ Creates visualizations, finds patterns, generate insights, and formulate predictive analysis from trade data using pandas, Quicksight, pyplot Show less

Philippines

- Creates Acceptance Criteria, Test Plans, and other documents for the whole process of SW development- Develops/Codes features committed for the sprint- Creates test cases, and tests modifications, unit tests- Handling and fixing of submitted prontos/bugs- Part of a Scrum Team- Expertise in C/C++ and Linux commands and Mobile Network architecture and protocols such as 4G/LTE- Uses tools such as Crucible, Git, Tortoise SVN, CLang, Jenkins, Gerrit, JIRA

Master Of Science - Ms, Computer Science

Major: Computer Vision and Machine Learning Published cited papers on different conferences and arXiV ○ Multi-class Semantic Segmentation.

Bachelor’S Degree, Computer Science

Activities and Societies: Society of Computer Scientist (SoComSci)

Education record

Activities and Societies: Engineering and Science Education Program (ESEP), MHPNHS Rondalla 13th Honorable Mention

Education record

Valedictorian