Milpitas, California, United States
I am passionate about tackling hard technical problems that are both important and motivated, leveraging a wealth of experience in a variety of computational paradigms such as big data, machine learning, artificial intelligence, search, and distributed computing.
• Research on event detection in text documents using a hybrid approach combining both traditional rule-based NLP and deep neural network approaches. • Research on applications of deep-learning-based computer vision in industrial inspection / supply chain digitization and frameworks for distributed AI model training. • Research focused on big data architectures for IOTs and applications of AI in the domain of renewable energy. • Winner of the 2017 European energy market forecast competition. Presented “Day-ahead wind power generation forecasting using Support Vector Machines” at the EEM 2017 conference in Germany.
• Hired to build the Intuit analytics cloud platform and develop big data analytics solutions. • Defined standards for the Intuit analytics cloud architecture and platform aimed to support building advanced analytics across various lines of business. • Developed a variety of clickstream data analytics solutions, including predictive models for conversion and product attachment that can be used as the basis for intervention and on-line marketing optimization. • Designed and implemented a platform-independent solution based-on PMML for automating real-time deployment of predictive models in big data environments, bridging an important gap between data science and engineering. • Designed and implemented a market sizing tool and an interactive query tool that helps generate complex analytics queries for efficient execution against very large databases, automating a manual research process that used to take weeks to complete. Both tools, metadata-driven and designed for use by domain experts and non-technical users, bridge another important gap between domain analysts and IT. • Prototyped real-time analytics solutions for fraud detection using scalable rules engines based on the RuleML standard. • Designed and implemented web-scale analytics using Map/Reduce and MPP relational engines, including distributed logistic regression. • Contributed to the design and implementation of the Intuit experimentation platform, an A/B testing platform which was later open-sourced as Wasabi. • Worked on automated extraction of mathematical formulas from instruction manuals provided in pdf. • Sample technologies used: Cassandra, Netezza, Spark, Kafka, Java, Python, AWS cloud.
• Delivered data warehousing and business intelligence solutions based on Neoview, a new MPP database platform developed and marketed by HP. • Optimized performance of Neoview applications often leading to orders of magnitude speedup. • Developed new product requirements for the Neoview R&D group based on requirements from customers and BI partners.
• Performed bid preparation and presentation, business development, requirement analysis, architectural design, detailed design, and implementation of substantial BI solutions for major HP customers. Technologies include dimensional modeling, OLAP, advanced reporting, enterprise data warehousing, MPP engines, and data mining. • Contributed to the design, implementation, validation, and on-time delivery of a substantial data warehouse/OLAP database system that facilitates real-time reporting and analysis of a comprehensive suite of key performance metrics for a major printer manufacturer.
• Conducted business development and market analysis in the space of music practice solutions. • Designed and implemented software-assisted methods for recording different instruments to separate channels, automated audio segmentation, and high-fidelity pitch-preserving time compression. • Launched Matching Tunes, the company’s first product for piano concertos practice.
• Architected, designed, and implemented the SurroMed’s integrated bioinformatics platform which was critical to the company’s competitiveness in the business of biomarker discovery. The platform made it 50x more efficient in the interpretation, storage, and analysis of biological measurement data, and led to the rapid discovery of SurroMed’s first biomarkers. The platform features a scientific OLAP system for bioinformatics, innovative clustering techniques to reduce data dimensionality, and the use of massively parallel predictive modeling on computer clusters. • Spearheaded the company’s data mining initiative and established the Surromed’s data mining group. Led the group to deliver scientist-oriented tools that are key to shorten the biomarker discovery cycle and to provide custom discovery services. Represented SurroMed through high-profile external presentations. Contributed to writing grant proposals and building external partnerships. • Directed the Bioinformatics department with a staff of four project managers and nine indirect reports, five of which have a Ph.D. Coordinated with other departments on setting objectives and priorities. Led department to deliver five robust software systems that are critical to SurroMed’s revenue-generating bioanalytical operations. As a member of the senior management team, contributed to business development and strategic planning. Advised executive management on key bioinformatics technologies.