Jake Ogle

Focus: ML systems · In progress (started 2026)

Develop and evaluate computer vision and large language models on large-scale biomedical datasets, including medical imaging and clinical text such as radiology reports. Lead the design, curation, and long-term stewardship of multimodal datasets spanning imaging, pathology, and report-level data. Build preprocessing pipelines for clinical imaging and text normalization, and maintain SQL-backed databases integrating heterogeneous clinical sources. Perform large-scale data processing and model training on Slurm-managed HPC clusters, with an emphasis on data quality assurance, versioning, and reproducible experimentation. Contribute to dissemination of research through peer-reviewed publications, conference proceedings, and academic presentations.

Designed and deployed production data and ML systems across real-time computer vision, LLM-backed retrieval, and telemetry pipelines. Built OpenCV-based vision pipelines for automated manufacturing monitoring; developed RAG systems using embeddings, vector databases, and LLM orchestration to reduce data access latency from minutes to seconds. Implemented centralized Python REST APIs integrating SQL Server and Azure SQL, and engineered scalable ingestion and validation pipelines with Apache Airflow. Contributed to hybrid cloud/on-prem architecture using Azure data and ML services, with a focus on observability, reliability, and long-term platform scalability.

Built automated build, test, and deployment pipelines using Python, Jenkins, and Groovy to support reproducible software delivery. Developed internal tooling for code signing and automated data collection, improving validation and traceability of software artifacts. Contributed to MISRA-compliant embedded C codebases and collaborated cross-functionally to ensure data integrity in production systems. Actively supported internal Python education initiatives to expand data processing capabilities across engineering teams.

Performed large-scale analytics on vehicle diagnostic datasets using SQL and cloud-based data platforms, identifying reliability and performance patterns in networked vehicle systems. Supported feature engineering and data preprocessing for ML workflows, enabling improved predictive diagnostics and system-level insight.

Engineered and managed clinical genomics data pipelines supporting rare disease research, including whole genome and whole exome sequencing (WGS/WES) projects under the Gabriella Miller Kids First initiative. Built and maintained research databases and patient tracking systems to support multi-institutional collaboration. Applied Python, R, and SQL for genomic data cleaning, normalization, and analysis, enabling reproducible downstream research workflows. Worked directly with sequencing data, metadata, and clinical cohorts, bridging clinical research requirements with scalable, computation-driven data management.