Unblocked Solutions LLC provides consulting with a no-nonsense, stepwise approach to getting your key enterprise information curated into forms best suited to leverage AI, retrieval based AI augmentation, and discrete planning systems. Unblocked Solutions LLC works with your domain experts and partners with 3rd party tools and platforms to get your AI transformation "unblocked".

Part of a team working on edge distributed systems to train large numbers of AI models for thousands of cell towers and radio components.

Built a simulation system for the HDS8000 data center computer product line, involving large-scale component configurations and use. Part of a team that built a distributed computing bootstrap system, including the implementation of HTTP-Signatures to secure GRPC calls, and a system for reversible workflow, capable of checkpoint recovery and cleanup of failed state transitions.

Developed a Node.js based system to automatically convert arbitrary JSON data to a new and efficient style of SQL normalization using power-law information for effective data compression. Deployed a Manta Map/Reduce ETL pipeline combining 10 separate JSON data sources with this tooling to create a data warehouse for internal IaaS data on a PostgreSQL 9.4b platform featuring automatically derived schema with an optimized mix of JSONB and SQL types. Provided complex SQL queries, analytics and views with R and Tableau on overall cloud performance and customer behavior.

Built demonstrations of Big Data computing in-situ on the Joyent Manta Storage Service. Strategic Planning and Competitor Analysis. Expanded the breadth of software support for Joyent's Manta Storage Service by building a Software Development Kit for the R language: https://github.com/joyent/mantaRSDK.

Managed the hiring, team organization, technical architecture, prototyping and deployment of the Mechanobiology Information Resource http://www.mechanobio.info/

Primary graduate supervisor of 5 Ph.D candidates in the Singapore-MIT Alliance program who were co-supervised by MIT faculty.

Academic teaching and research duties. Taught in multiple courses in Bioinformatics spanning topics from Molecular Evolution, Proteomics, Structural Biology at both the graduate and undergraduate levels. Supervised 6 Ph.D., 2 M.Sc. and 8 undergraduate thesis project students.

Supervised 6 graduate students, and carried out core Bioinformatics undergraduate course teaching duties, as well as undergraduate research project supervision. Status-only appointment, while employed at Mount Sinai Hospital.

Organized and ran the first Bioinformatics training workshops at the National University of Singapore, exporting coursework and expertise gained from the Canadian Bioinformatics Workshops.

Spun out of Mount Sinai Hospital and backed with $1M in hardware contributed by Sun Microsystems, Unleashed Informatics's first product was DogBox. DogBox was an advanced implemention of the SeqHound integrated database API server. The product offering was a $28,500 standalone Solaris 10 x86 based Sun v20z server with StorEdge FC3511 storage array. Software included 18 integrated public life-science databases and API services for PERL, Java, C and C++. Daily data updates were provided by an SSL-based InnoDB/MySQL based remote master-slave update mechanism, SecurSync, which was able to transfer large updates over low-bandwidth, high-latency connections. Software and security patches were also managed remotely.In late 2005 Unleashed Informatics obtained the commercial rights to BIND and related tools including the 2.3 million-record Small Molecule Interaction Database (SMID). The platform was renamed BOND (Biomolecular Object Network Database) and delivered browser, REST and SOAP API access to molecular interactions, 3D structures and reference protein and nucleic acid sequences. An Apache Lucene-based text index provided a comprehensive index of contents. NBLAST provided a complete set of precomputed BLAST neighbors for the nr protein database.In March 2007, Unleashed Informatics achieved a successful commercial exit through sale to Thomson Scientific, a division of the Canadian company Thomson Inc. (now Thomson-Reuters), and used to support expansion of the GENESEQ patent sequence database with public reference information and annotation.

Corporate entity established to carry out remote BIND Database curation funded by the Economic Development Board of Singapore, and partnered with the National University of Singapore and the Genome Institute of Singapore.

The Blueprint Intitiative's primary mission was the development of the first large-scale molecular interaction database, BIND, which was based on NCBI Asn.1 C toolkit and standards as developed in the Ph.D. thesis work of Gary Bader. BIND 3.0 was released in June 2003 rebuilt on XML and JAVA. The first documented curation practice and training materials in the public Bioinformatics sphere were released in 2003. Blueprint WorldWide Inc. was established by MDS-Proteomics and IBM in 2001 as a not-for-profit company to raise funds for developing the public open-acces BIND database, together with related tools and the SeqHound API service. Funding from Genome Canada and the Ontario R&D Challenge fund was secured by April 2002. In May 2002 management and governance was shifted back to Mount Sinai Hospital's Samuel Lunenfeld Research Institute's Blueprint Initiative, under the leadership of Dr. Hogue. Governance gridlock led to the replacement of in-kind contributions from MDS-Proteomics and IBM with Sun Microsystems in November 2002.A total of 211 person-years of database curation and software development effort led to a total of 208,173 interaction database records spanning all organisms. Journal relationships for pre-publication database curation were established with Nature, Science and other leading journals.In late 2005. software intellectual property was retained by Mount Sinai Hospital and licensed for commercialization to Unleashed Informatics. The Blueprint Initiative was Genome Canada's largest investment in a Bioinformatics project and the first Genome Canada effort that had a successful commercial exit to a Canadian company. It was also the first tenant of the Toronto MaRS incubator. The original public BIND data has been reformatted to PSI-MI format by the Bader lab in 2010, and continues to serve scientists worldwide.

MDS Proteomics resulted from a merger of Ocata Proteomics Inc and Protana. MDS Proteomics industrialized high-throughput LC-MS/MS Mass Spectrometry, using the MDS-Sciex and Thermo platform instruments for the purpose of protein function discovery.CIO Roles and Responsibilities:Responsible authority for company-wide informaticsEstablish and achieve the goals for deploying IT for business valueCreate the blueprint for enterprise-wide IT strategy and standardsMaximize return for every dollar spent on ITDetermine the short list of vendors capable of delivering the best solutionsApprove spending for all IT-related products and servicesConvert data and information into knowledgeEmpower the organization to meet its customer's demandsSupport ongoing competitive positioningFacilitate internal teamwork and external partnershipsSoftware ArchitectStrategic decision maker regarding the design and implementation of in-house and third party software integration.Ensure the software development environment suits the selection of tools and methodologies.

Under the guidance of Stephen H. Bryant, undertook the development of an API for 3D molecular structure applications (MMDB-API) based on the Asn.1 MMDB schema. Subsequently developed Cn3D 1.0 and 1.1 together with Alex Smirnov and Denis Vakatov who contributed the platform-independent Vibrant extensions required for 3D graphic display. Released the first NCBI 3D visualization service in 1995 supporting RasMol and automatically rendered MAGE images. Cn3D 1.0 was released in 1996 supporting Windows, Linux, Solaris, Irix, and DEC OSF/1. Awarded "Best of Web, Chemistry, 1996", (http://www.ch.ic.ac.uk/rzepa/RSC/chemsocrev/). Cn3D 1.1 supported 68K Macintosh platforms.