Miguel Moisés Serrano, Ph.D. Email & Phone Number

United States

Atlanta, Georgia, United States

Lead ML Engineer:* Creating, Developing and Deploying ML solutions.

San Francisco Bay Area

* Led the development of a pre-consult summary generator using large language models (LLMs) to extract and summarize comprehensive patient information, reducing provider preparation time.* Collaborated on a chat response generator that utilized LLMs, significantly reducing provider billable time by automating initial communication with members.* Created models leveraging medical, pharmaceutical, social determinants of health, and demographic attributes to enhance call center candidate identification, using explainability metrics to accurately extract attribution for the campaign’s performance.* Collaborated on a company-wide project automating member assignment to health programs based on medical, pharmaceutical, and demographic attributes, predicting outcomes and impact on revenue and cost.* Managed model development projects, ensuring clear communication, reducing duplication of work, and fostering mutual gains across teams.* Drafted Technical Review Documents outlining solutions and stakeholder information, including use cases, benefits, and key performance indicators (KPIs).* Collaborated with multiple departments, meeting diverse data science requirements throughout the organization.* Designed MCMC simulations to analyze contractual commitments' impact on revenue/cost, optimizing outreach strategies on a per-contract level.* Connected model performance metrics to financial impact, demonstrating the value of machine learning solutions.* Programmed package that utilized gcp’s big query ml to train and deploy models for the data analyst team, removing the barrier to entry and allowing them to building production ready models * Supported various efforts to clean, wrangle, acquire and serve data

Marietta, Georgia, United States

* Designed, developed and launched a Cash Flow Forecast tool that is currently used to guide business decisions and initiatives* Led multiple projects resulting in the development of products and contributed significantly to three patent submissions currently under review.* Led the team in adopting best practices for cloud infrastructure, code management, and CI/CD, resulting in improved job management, development processes, and product maintenance.* Mentored an intern and guided their efforts in creating an automated call center staffing solution.* Served as the subject matter expert in data science, specifically focused on machine learning and deep learning frameworks for relevant projects.* Designed and implemented a cloud-based infrastructure utilizing Docker, Kubernetes, and Airflow for efficient job management and product deployment.* Leveraged Jupyter notebooks, GitHub, and MLflow for experiment management within the cloud-based infrastructure.* Utilized machine learning tools to extract insights from Revenue Cycle Management metrics.* Developed various models to estimate client-based performance metric projections.* Created the Cash Flow Forecast, a per-client claim-level prediction solution, and implemented the cloud infrastructure to support it.* Produced a predictor for per-instance net collections over time profiles, enabling manager-level drill-down options to identify cash flow opportunities.

Greater Atlanta Area

* Managed the efforts of interns, developers, and fellow machine learning engineers to successfully deliver a model-agnostic cloud-based training, evaluation, and inferencing pipeline for aerial imagery.* Oversaw collaborations with research partners at GTRI, facilitating communication, updates, and objective establishment.* Applied modern machine learning methods to detect various asset types from aerial imagery, aligning with business objectives.* Established, coordinated, and managed the annotation team, which processed approximately 3,200 annotations per day.* Defined machine learning solutions to meet current business needs and objectives.* Designed a semantic segmentation model using a variant of the Segnet framework for land use analysis.* Utilized WMS, XYZ, and WMTS mechanisms for efficient and consistent data access and dataset generation.* Integrated research toolkit with the development environment for faster access to data as it became available.* Coordinated and supported external machine learning partners.* Completed leadership training courses provided by Airbus.* Built an end-to-end training pipeline using Docker, Kubernetes, and Airflow for complete process orchestration.* Dockerized training pipeline components, including sample set generation, training, validation, and prediction.* Utilized Jenkins for maintaining each model repository, ensuring CICD standards were met.* Designed the final component of the training pipeline to generate multiple Docker containers, each supporting different types of services with the same model.

Atlanta, Georgia, United States

* Investigated the use of hybrid machine learning and model-based computer vision for human pose estimation, focusing on practical methods for medical applications.* Utilized de-convolutional neural networks to identify subsections of subjects (e.g., human, infant, robotic manipulator) for accurate human pose estimation.* Incorporated model-based methods for synthetic training set generation, contributing to realistic and practical semantic segmentation.* Leveraged articulated models to create robust features for training a 3D pose estimation predictor.* Designed and currently developing a protocol for training and testing a robotic manipulator pose estimation method, providing real-time feedback in a real-world pick and place application.* Created a marker tracking tool for tracking lead markers on subsurface undulating animals in sand using X-ray images.* Applied pose estimates in both 2D and 3D for clinical gait analysis and infant kicking health analysis, respectively.* Developed an Institutional Review Board (IRB) protocol to enable the study and design of tracking for infant kicking analysis.* Provided guidance and expertise in a project involving deep convolutional regression for converting range image captures and goal directions into optimal path decisions for a planar robot.* Conducted research on topics including biomimetics, vision controls, navigation, SLAM (Simultaneous * * Localization and Mapping), vision-based pose estimation for physical therapy applications, construction robotics, and machine learning methods for object detection and tracking.* Mentored undergraduate and graduate students during their transition from pure studies to research.* Managed multiple undergraduate teams in various research programs across the aforementioned fields of study.* Created research-worthy lines of inquiry and delivered weekly reports and presentations outlining progress made.

Instructor of record for an intro to electrical engineering course for two consecutive semesters. Additionally, taught a series of lectures in controls, robotics, and computer vision course.

Atlanta, Georgia

* Conducted research and developed methods to detect and categorize unique structures from densely reconstructed point cloud data, with the output converted to industry-standard formats such as CAD.* Utilized machine learning methods, including deep convolutional neural networks and support vector machines, to extract semantic and quantitative values from point cloud reconstructions and aerial drone images.* Developed a semantic segmentation protocol using TensorFlow to detect and isolate structures from drone aerial imagery.* Applied clustering methods to isolate geometric primitives in sparse point clouds.* Estimated indoor room boundaries based on granular dense reconstruction from video capture.* Designed and implemented an innovative active polygon algorithm segmentation, similar to active contours.* Devised a metric system for evaluating the performance of roof estimation algorithms.* Created a nearest neighbor-based polygon filter for further refinement and analysis.

Doctor Of Philosophy - Phd, Electrical And Electronics Engineering

Master'S Degree, Electrical And Electronics Engineering

Activities and Societies: Controls, Computer Vision, Robotics, and Machine Learning

Bachelor'S Degree, Electrical And Electronics Engineering

Activities and Societies: Eta Kappa Nu, Tau Beta Pi