CONSULTING

  1. Research and Data Consulting: 

    We provide expert guidance on acquiring, storing, managing, analyzing, and using biomedical data for research. Consultations help investigators assess feasibility, define technical requirements, estimate effort and resources, identify appropriate methods, and determine when additional MIAI Core or BMI faculty expertise is needed. Support is available throughout the project lifecycle, from early planning to implementation and troubleshooting.

  2. CohortCount: 

    We estimate the number of patients within Emory Healthcare and affiliated facilities who meet defined eligibility criteria. This service helps investigators determine whether a sufficiently large study population is available before beginning a study, preparing a grant, or requesting a full data extraction.

  3. Grant and Protocol Support:

    We provide tailored technical language and letters of support for grant proposals, IRB protocols, project reports, and data management plans. Content may address data acquisition, processing, integration, infrastructure, security, statistical analysis, machine learning, and related biomedical informatics methods to support grant applications.

  4. Project Management:

    We provide project planning, coordination, monitoring, and reporting for biomedical informatics and AI projects. Services may include defining scope, milestones, responsibilities, dependencies, deliverables, timelines, and resource needs; coordinating core and external activities; tracking progress; and addressing project risks. Support is available from project conceptualization through completion.



DATA EXTRACTION

  1. Clinical Data Extraction:

    We extract clinical data from electronic health records, bedside monitor waveform data and related institutional systems according to project-specific requirements. Data may include structured records, clinical notes, laboratory results, medications, diagnoses, procedures, physiological time series, medical images, videos, and other research-relevant information.

  2. External and Device Data Acquisition:

    We acquire and integrate data from sources outside standard clinical databases, including wearable devices, sensors, mobile applications, public databases, social media, and other approved external resources.

  3. Data Integration and Preparation:

    We combine data from multiple sources and prepare it for analysis. Services may include data normalization, ontology mapping, temporal alignment, quality assessment, cleaning, de-identification, restructuring, and conversion into analysis-ready formats.

  4. Streaming and Near-Real-Time Data:

    We develop continuous or scheduled data feeds for projects that require current clinical, sensor, or operational data. This service supports near-real-time monitoring, dashboards, model inference, dynamic prediction, and continuously updated research workflows.

  5. Data Exploration and Feasibility Analysis:

    We perform focused exploratory analyses to assess data availability, quality, completeness, distributions, and suitability for a proposed research question. This service may include sample data pulls, summary statistics, visualization, preliminary signal processing, and evaluation of candidate analytical methods.

  6. Visual Analytics and Dashboards:

    We develop interactive dashboards and visual analytics tools that help investigators explore complex datasets, monitor project measures, review model outputs, and support data-driven decisions.

  7. Machine Learning and Statistical Analysis:

    We develop and apply statistical, signal-processing, machine-learning, and deep-learning methods to biomedical data. Services may include feature and biomarker extraction, model training and evaluation, classification, prediction, clustering, time-series analysis, image analysis, and multimodal data analysis.

  8. NLP and LLM Analysis:

    We analyze clinical notes and other text using natural language processing (NLP) and large language models (LLMs). Services may include information extraction, named-entity recognition, automated annotation, cohort discovery, text classification, de-identification, lexicon expansion, retrieval-augmented generation, prompting strategies, language-model adaptation, and end-to-end clinical text-processing pipelines.

     



SOFTWARE AND MODEL DEVELOPMENT

  1. Application Development:

    We design and develop web, mobile, desktop, and backend applications for biomedical research. Applications may support data collection, annotation, visualization, analysis, workflow management, decision support, and real-time monitoring. Services may include user-interface design, database integration, authentication, access control, and application programming interfaces.

  2. Data and Software Pipeline Development:

    We develop reproducible software pipelines for data acquisition, preprocessing, quality control, integration, analysis, visualization, and reporting. Pipelines may be designed for batch processing, high-performance computing, cloud environments, or near-real-time operation.

  3. Machine Learning and AI Model Development:

    We design, train, validate, and test machine-learning and deep-learning models for project-specific research questions. Services may include feature engineering, model selection, hyperparameter tuning, cross-validation, uncertainty estimation, interpretability analysis, and comparison of alternative methods.

  4. NLP and LLM Model Development:

    We develop and customize natural language processing and fine-tune small/large language models for biomedical applications. Services may include prompt development, retrieval-augmented generation, supervised model training, fine-tuning, domain adaptation, automated annotation, information extraction, text classification, de-identification, and customized language-model pipelines.

  5. Model Optimization and Transfer Learning:

    We improve the computational efficiency and performance of existing machine-learning, deep-learning, NLP, and LLM models. Services may include transfer learning, model adaptation, quantization, pruning, memory optimization, distributed computing, inference acceleration, and reduction of cloud or on-premises computing costs.

  6. Synthetic Data and Digital Twins:

    We develop synthetic datasets, simulation models, and digital twins that reproduce selected properties of real biomedical systems or patient populations. These tools may support model development, data augmentation, scenario testing, rare-event simulation, class balancing, algorithm evaluation, and data-sharing workflows when access to real data is limited.



SOFTWARE AND MODEL DEVELOPMENT

  1. Application and Software Development:

    We deploy research applications, software tools, databases, dashboards, and analytical pipelines into approved operational environments. Deployment may occur on institutional servers, virtual machines, high-performance computing systems, or cloud infrastructure.

  2. Machine Learning Model Development:

    We convert trained machine-learning and deep-learning models into operational services that can process new data and return predictions, scores, or other outputs. Services may include model packaging (containerization), application programming interfaces, dashboard integration, workflow integration, performance monitoring, and version control.

  3. NLP and LLM Development:

    We deploy NLP and pretrained/fine-tuned LLM systems for research and operational use. Services may include model hosting, retrieval systems, vector databases, prompt management, access controls, logging, monitoring, model optimization, and integration with clinical or research applications.

  4. Software and Model Validation:

    We evaluate software and computational models using rigorous and reproducible training, validation, and testing procedures. Validation may assess accuracy, generalizability, robustness, reliability, bias, calibration, uncertainty, computational performance, and performance on independent or previously unseen data.

    When appropriate, we also support investigators and external partners preparing evidence for regulatory review, including the evaluation of software and models using standardized datasets and documented validation procedures.

  5. Data Management:

    We organize, store, maintain, and govern identified, limited, and de-identified clinical and research data. Services may include database and data-lake design, metadata management, access controls, encryption, backup, data-quality monitoring, archival, retention, deletion, and preparation for submission to approved public repositories.

  6. Infrastructure Maintenance and Operational Support:

    We provide ongoing support for deployed applications, software, models, databases, and computing environments. Services may include system monitoring, updates, security patches, troubleshooting, backup, performance optimization, user support, and coordination with institutional IT and security teams.



COMPUTER AND STORAGE

  1. CPU Computing:

    We provide access to CPU-based computing resources for data processing, statistical analysis, software development, and model training or inference that does not require specialized accelerators.

  2. GPU Computing:

    We provide GPU-based computing resources for machine learning, deep learning, image processing, signal processing, and other computationally intensive workloads.

  3. LLM Computing:

    We provide specialized computing resources for training, fine-tuning, evaluating, and deploying large language models and other high-memory AI workloads on LLM-specialized machines.

  4. Research Data Storage:

    We provide secure storage for clinical, research, and computational data using approved on-premises and cloud infrastructure.

  5. Virtual Machine Hosting:

    We provide virtual machines for hosting research applications, databases, dashboards, software tools, and model-serving environments.

    The base configuration includes:

    • 100 GiB of storage
    • 4 CPU cores
    • 4 GiB of RAM
  6. Infrastructure Support:

    We provide technical support for computing, storage, networking, data transfer, access control, system integration, and research infrastructure. This may include configuration, maintenance, troubleshooting, performance monitoring, and coordination with institutional IT services.