IT Director, Imaging Center — Sr. Manager, Bioinformatics — Associate in Psychiatry

McLean Hospital — Mass General Brigham — Harvard Medical School 2019 - Present

Belmont, MA

• Set IT strategy and budgeting across clinical and research programs for the McLean Imaging Center; set mission, roadmap, and capacity plans for compute, storage, and networking.
• Architect resilient infrastructure with operational continuity; lead purchasing, capacity management, and lifecycle planning.
• Deliver enterprise backup & DR with on- and off-site replication supporting disk, machine, and site-level recovery.
• Build observability and QA: Zabbix-based real-time monitoring/alerting and trend-driven predictive models; performance metrics to forecast growth and utilization.
• Deploy modern imaging platforms including XNAT and Orthanc on TrueNAS + Proxmox; reproducible configuration-as-code; lossless ETL of images, metadata, and users.
• Develop reproducible neuroimaging pipelines: automated fMRIPrep workflows driven by declarative YAML.
• Maintain research computing: Red Hat + Bright/SLURM HPC; current imaging/system software; shared storage, backup, and networking for Core and Clinical Services; manage PACS and RIS.
• Virtualize services to reduce hardware footprint and improve agility and reliability.
• Direct desktop support operations: standards, processes, training; recruit, develop, and retain a high-performing team.
• Bridge IT and neuroscience as grant co-investigator: shape experimental design, build analysis software and data flows, interpret results, and contribute to publications; collaborate with external scientists.
• Partner with clinical and research stakeholders to refine workflows; evaluate and adopt emerging IT, data science, and imaging technologies to meet evolving community needs.

• Supported clients, including many of the Fortune 50, in their use of the Ab Initio ETL software including the GDE, Co>Operating System, Control Center, Metadata Hub, and underlying systems.
• Built software solutions on Linux, Solaris, and AIX ranging from low-level system utilities to high-performance data processing engines operating over data warehouse systems spanning tens of thousands of database tables and hundreds of petabytes of data.
• Implemented systems to provide data lineage traceability and build systems that supply reproducible reporting.
• Taught internal classes in the use of our technologies and developed additional teaching materials to enhance existing classes.
• Managed and helped develop the internal Help system, which utilized open-source components such as nginx and Tomcat.
• Developed and maintained internal-facing dashboards (web-based and using custom designed hardware based on the Raspberry Pi and Arduino platforms) to support visualization and escalation.
• Started and managed a lecture series, "Brainfood", wherein colleagues presented interesting scientific concepts, their thesis work, or anything else they found fascinating.

• Developed software (Python, bash, Tcl/Tk, and C) for FDA and CE certified robots that assist patients with rehabilitation.
• Worked with clinical staff to design treatment protocols and translate these into products.
• Made sure the robots followed Asimov's First Law of Robotics by implementing multiple redundant safety systems including software- and hardware-based power-off systems triggering on logical, electrical, and mechanical faults, with hard real time (<1 ms) decision time requirements.
• Rewrote tens of thousands of lines of legacy Tcl/Tk code in (much less) Python.
• Developed software in C for an Atmel microcontroller, replacing the PLC in a new product line.
• Developed robot mechanical dynamics simulations in Jupyter/IPython Notebook.
• Rewrote the central Xenomai-based real-time robot control loop to use a CANbus architecture.
• Redesigned software deployment mechanism, which had been based mostly on manual instructions and loose collections of shell scripts, to use modern Debian packages and containerization.
• Mentored junior programmers and interns.

• Heuristic analysis, usability testing, statistics, data analysis, and background research in the service of medical device usability and safety, including human factors testing for regulatory approval.
• Applied principles of human psychology and cognitive performance to improving usability.

At these three positions:

• Developed tools in Perl and shell to monitor Unix systems and networks.
• Supported Unix systems providing financial and other messaging services to 100,000+ customers in a high availability environment.
• Created formal documentation of procedures that had previously been passed down orally.

Thesis: Neural Object Representation Spaces and their Metric Properties

• I showed a new method of computationally modeling and analyzing functional neuroimaging data using multidimensional scaling-like methods, and used it to differentiate between distributed and localized neural encodings.
• I wrote tools in Python and Matlab to manage long-running computation jobs and the operation of the Penn Center for Functional Neuroimaging cluster, and created a calendaring system to manage MRI reservations with complex business process requirements. I additionally took coursework in the theory of human decision-making: the science of how and why people and organizations make mistakes, and how to overcome these mistakes and make rational decisions instead.