Instructor in Machine Learning and Artificial Intelligence
2016 - Present
Having created Harvard's first survey course on the history and current state of machine learning and artificial intelligence, I continue to offer the course one to two semesters each year. Students learn using lectures, readings, flipped-classroom exercises, collaborative on-line coding (e.g., using Jupyter Notebooks and Google Colab) and some hands-on extracurricular activities. Each semester requires updating approximately one-third of lecture slides and supplemental materials. As a higher-level course suitable for non-CS students as well, we cover neuroanatomy, neurophysiology, machine learning, data science, neural networks, artificial intelligence and many of the socioeconomic, ethical, legal, and political topics surrounding data science and AI. With colleagues in the Boston technology sector (including the vibrant venture capital sector) and at the Division of Continuing Education (Extension School), I am working on a proposal for a technology leadership and entrepreneurship certificate program.
Researcher in Neuroinformatics
2014 - 2020
My research appointment permits me to continue to advise and collaborate with colleagues and peers at the university as a researcher. Along with my instructor's appointment, I am able to continue teaching and doing some research while I pursue other endeavors. My research interests are in information technology leadership and the application of computer science, visualization, and data science in the biotech, healthcare, and higher education space.
Co-Chair, University IT Talent Acquisition, Diversity, and Retention Committee
2016 - 2017
Higher education is generally unable to match many of the private sector total compensation packages for computer-related positions. In spite of the otherwise exceptional benefits offered by higher education, salaries as well as other benefits are often significantly higher in the private sector, at times by more than 75% for the top performers. Unfortunately, the allure of working for highly visible companies like Google, Facebook, Amazon, and Microsoft exceeds even the prestige of Ivy Plus universities. Leveraging an aggressive diversity and inclusion strategy which includes forming ties with other institutions traditionally associated with underserved communities is not only a moral imperative but a survival and value-proposition strategy. Developing and retaining talent is also an essential competitive strategy for attracting and retaining top talent. As one of two founding Co-Chairs of the University IT Talent Acquisition, Diversity, and Retention Committee, we focused on building and maintaining strong talent pipelines that increased our diversity numbers as well as devising methods to engage, grow, and retain existing exceptional talent.
CIO Sponsor & Committee Chair, University Access to Institutional Data Strategic Initiative
2015 - 2017
Access to institutional data is of growing importance to all. In higher education, it is pivotal to accomplishing the core mission of the institution while protecting privacy and academic freedom. Harvard realized the importance of access to institutional data; however, security, privacy, data integrity, and accountability were and remain of paramount importance. At the request of the University Vice President and CIO, I lead the initiative for the duration of my tenure as the Asst. Dean for Computing and CIO of the School of Engineering and Applied Sciences. The committee set forth a specific strategy and stated goals to permit increased access to institutional data to approved individuals across multiple disjointed systems thereby enabling evidence-based decision making from admissions to resource allocation and from alumni relationship management to cost optimization.
Co-Chair, University Research Data Initiative
2015 - 2017
With nearly 90,000 active users including affiliated researchers, enabling the twelve degree-granting schools and multiple affiliated institutes and hospitals to manage and steward research data is an enormous effort. As co-Chair of the University Research Data Initiative, I and my peer created and oversaw a strategic plan and goals for creating more uniform policies, services, and platforms for the diverse community needs. Working closely with University leadership in the Vice Provost for Research's office, the Faculty of Arts and Sciences Research Computing Group, the University's Chief Information Security Officer, and Harvard University IT, we created a framework for accelerating research, increasing reproducibility, and meeting or exceeding regulatory requirements up to and including FISMA Medium and NIST800-171.
Council Member, University CIO Council (Ex Officio)
2013 - 2017
As CIO of the Harvard John A. Paulson School of Engineering and Applied Sciences, I, along with the CIOs of all other schools, the head of Libraries, the University CIO and CISO sat on the Harvard University CIO Council where we were often able to reach consensus on a number of topics, policies, and strategic initiatives. Select endeavors include:
- Teaching and Learning.
- Teaching and Learning Technologies (TLT).
- Data for Learning Analytics.
- Learning Management System Migration and Integration (to Canvas).
- Online Access to Scholarship and Knowledge.
- Big Data for Research (Infrastructure & IT Organization).
- Student Information Systems Upgrade ($52M).
- Access to Institutional Data and Systems Integration.
- Common Platforms.
- Video (Live Streaming, Accessibility, and Retention).
- Unified Communication (aka “Harvard Phone”).
- Collaboration Tools (Document Shares, Co-Authorship, Chat, Calendar & Email).
- Faculty & Student Computing.
- Research Computing.
- SaaS, PaaS, IaaS.
- DevOps transformation.
- Identity and Access Management.
- Foundational Strategies.
- Enterprise Architecture and Service Bus.
- Information Security.
- Business Continuity and Disaster Recovery (BCDR).
- UX, Mobile, BYOD, and Accessibility.
- Vendor Management.
Member, Faculty IT Steering Committee
2013 - 2017
As a voting member of the FAS Faculty IT Steering Committee which oversaw the University-Wide Harvard University IT as well as all IT related items affecting Harvard College, the Graduate School of Arts and Sciences, the Faculty of Arts and Sciences, and the University Library System, the Committee members and I were responsible for providing oversight and guidance for many of the University's strategic plans and initiatives concerning information technology and computing. During my tenure, we oversaw the creation of the “One Harvard” single sign-on enabling approximately 140,000 active users to use one username and password to access all major systems and services. Other endeavors include:
- A $52M student information system transformation.
- Upgrading of Oracle Financials and transition from cash-based to GAAP accounting practices.
- Creation of a centralized procurement system with searchable contracts and agreements.
- Deployment of the “PI Dashboard” which my group created to all of the University.
- The implementation of cybersecurity best practices, including a simple “Data Levels” system for the community and roll-out of LastPass password manager.
Inaugural Asst. Dean for Computing & CIO
2013 - 2017
Harvard’s newest school, having had a half-dozen people head some incarnation of the IT and Computing groups in the prior five years, was in need of significant transformational change. I was selected in no small part due to my reputation for building strong, outperforming groups. Within three years, I had rebuilt the IT infrastructure and computing group while merging most commodity services collaboratively with Harvard University’s central IT organization (HUIT). I created a world-class team and a modern Cloud hybrid infrastructure and services organization. I decommissioned and merged multiple high-performance compute clusters into the Massachusetts Green High-Performance Computing Center to create the 5th
largest higher education compute cluster in the world.
Select accomplishments include:
- Created a multi-year strategic plan and mission, vision, values statement in close collaboration with the administration, faculty, researchers, and students.
- Reigned in costs while significantly increasing resources and services
- Reduced total IT spend, whether by the core budget or by faculty and community members, by 27%
- Reduced per-capita spends across all groups by 31% to 44%
- 83% of faculty said they were no worse off, 63% said we were better or much better off after reductions
- Collaboratively created 5th largest High-Performance Computing environment globally in higher education
- Increased available computing compute power nearly two orders of magnitude
- Increased researcher utilization from ~13% to over 85%
- Partnered with Amazon Web Services to make SEAS the world leader in usage of AWS Cloud computing in the classroom
- No cost to SEAS
- AWS provided engineers, infrastructure, and computing credits
- Pioneered collaborative cloud computing infrastructure
- Created first elastic implementation of JupyterHub currently being adopted at UC Berkeley, Cornell, and Yale
- The first school to move all infrastructure to the Cloud
- Realized $1.6M/year ongoing savings compared to original costs, while expanding services
- Enabled security and compliance auditing previously not possible
- Significantly increase infrastructure reliability by one order of magnitude (1.4×10-2 to 2.1×10-3 for downtime risk)
- Increased data durability from approximately 99.5% to 99.92% (99.9999999% or “10 nines” for fully cloud-based storage)
- Renegotiated all computing and IT contracts (hardware, software, and services) resulting in an average of 16% savings across all IT spend
- Enterprise Systems Mergers and Rollouts (with reduced spending every year, year over year)
- Merged 7 email systems into the two university-wide Google Education and Office365 systems.
- Merged two active directories and one LDAP server into one university-wide single AD
- Merged two legacy SharePoint servers and multiple SMB file servers into Office365 OneDrive
- Moved multiple SMB file servers and Google Drive for Education
- Rolled out two-factor authentication
- Rolled out CrowdStrike, BitLocker, AirWatch, and Bit9 for security
: retrospective report available in redacted format upon request.
Chair, IT Steering Committee
2013 - 2017
As the STEM school of Harvard University, the School of Engineering and Applied Sciences presented both the challenges and opportunities of serving and engaging world-leaders in computer science and electrical engineering. The IT Steering committee was devised from inception to provide reasonable guidance and oversight of all computing-related activities at the nascent school. With the understanding that, although desirable, unanimity could not be a requirement in decision making, I was able to affect enormous changes in a short period of time with surprisingly high levels of community satisfaction thanks in part to the constant, close engagement of faculty members on the IT Steering Committee.
Harvard asked me to take on additional roles at the Harvard Medical School affiliated Massachusetts General Hospital (MGH) in a 60%/40% split (Harvard/MGH respectively) during which I managed the high-performance computing and clinical research informatics at the MGH/HTS Athinoula A. Martinos Center for Biomedical Imaging as well as the Psychiatry and Neuroscience departments. Serving over 1,114 active clinicians and researchers across a half-dozen campus locations, we deployed a multi-petabyte extreme performance computing infrastructure with over 2,000 nodes with over 8,000 CPU and 80,000 GPU cores enabling the computing required for the Human Connectome Project from the functional MRI to sub-nanometer transmission electron microscopy resolution. The algorithms, services, and infrastructure enabled best-in-class bleeding-edge research in medical imaging including fictional connectivity, tractography (e.g., diffusion imaging), and machine-learning image recognition.
Much of my time as Head of Neuroinformatics was concurrent with my role as Director of Clinical Research Informatics at MGH. As Head of Neuroinformatics, I created and directed the groups responsible for the architecture, implementation, and support of the computational infrastructure supporting over 26 Harvard faculty, over 200 active clinical and non-clinical researchers, and over 60 major projects.
Select accomplishments include:
- Accomplished objectives of the US Presidential “Functional” Human Connectome Project for 90% less than projected ($30M vs. $2.7M) and with 4,600 high-resolution human subject scans (2,000 original targets).
- Took a non-existent group, with no formal neuroscience training, to an internationally recognized leader in neuroinformatics, while keeping the group to 1/10 the size of our most significant competitor (WUSTL).
- Created automated scientific pipelines adopted across multiple universities and dozens of research hospitals nationally and internationally.
- Partnered closely with world-leading faculty in neuroscience eventually having almost all faculty use the high-performance automated processing pipelines and tools.
- Increased the reliability, usability, and reproducibility of raw acquired neuroimaging data resulting in a reduction from approximately 23% to 4% “unusable” or “noisy” scans.
- Architected and prototyped high-performance computing infrastructure later adopted broadly by Partner’s Healthcare hospitals and Harvard University Faculty of Arts and Sciences. Several researchers at other universities including UC Berkley, Yale, MIT, Tufts, and the University of Virginia also adopted several of our tools and methodologies. This included the encapsulation of “pipeline modules” into light-weight portable virtual machines that functioned much like cloud containers prior to their invention.
Co-Creator of the interdisciplinary Neuroinformatics project which eventually became the Neuroinformatics Research Group at Harvard. Harvard's short-lived but exciting “Initiative in Innovative Computing” was a platform to launch new initiatives at Harvard across interdisciplinary fields. I, along with two others, submitted a proposal to create an infrastructure and coding platform to allow the still relatively emergent field of neuroinformatics to leverage advances from many different fields from IT to radio astronomy to further our understanding of the human brain.