Effectivizing product support through machine learning and pragmatic algorithmic or data-driven solutions.Analysis of conversations to produce 80/20 actionable suggestions on impactful content.Content generation through GPT3, natural-language reasoning, and a host of reliability improvements including an in-house structured optimization approach. On a particular key dataset of real open-ended customer queries, we produced 97 % accurate answers. Web and document sources.Intent classification and other forms of NLU with modern transformers such as DeBERTa, efficient trade offs, and combinations of strengths from traditional and modern approaches. On a particular dataset of real customer queries, beating all competitors in understanding intent including 41 % error reduction viz-a-viz GPT3.Knowledge representation facilitating generalization across situations and product families while checking boxes of being intuitive, controllable, suitable for optimization, exploration, and learning from data.Productionalization of AI/ML solutions as services through Docker and GCP; negotiation of integrations.Prototyping, ideation, design, and prioritization with the product team.Good directions, plenty of practical quirks to address still.

Projects:* Small-scale machine-learning SaaS developed from idea inception to deployment and management. Involved pilot studies to determine feasibility, solution architecture, API design, integration, development * Web platform to support integration and development speed of machine-learning solutions. Development and business planning with various models. Involved understanding the broader picture of integration, data analysis, cloud services, machine learning, web technologies, and user interfaces. Did not get past early investigations.#Data Science, #Machine Learning, #Solution Architecture, #Business Development, #SaaS, #Data Analysis, #Docker, #API Design, #REST, #scikit-learn, #pandas, #Python, #R, #NodeJS, #MongoDB, #Deep Learning, #Integration, #Fintech, #Agile, #Lean, #Amazon Web Services, #Market Analysis, #Microservice, #TensorFlow, #Big Data, #Frontend, #UX, #Git, #GraphLab, #Hadoop, #Spark, #Microsoft Azure

Infectious-disease modelingDeveloping and analyzing models for influenza pandemics and transfusion-transmitted diseases. Evaluating the impact of public interventions and integrating targeted domain knowledge. Specifically, adapted simulations to new scenarios and interventions rooted in literature, overhauled simulations to improve development speed, refactored I/O data dependencies, optimized overall system performance by a factor about 40, and performed exploratory data analysis and presentation. Exploratory analysis of predictive models and anomaly detection from web data.#Modelling #Statistics #Simulation #Data Analysis #Data Science, #Algorithms #C++ #Ruby #R #d3js #Statistics #Monte Carlo #Public Policy #Healthcare #Optimization #boost #ggplot #Shiny #Tidyverse #Epidemiology #Infectious Disease #Predictive Modelling #Visualization #Influenza #Pandemics #Virology

Analyzing, designing, and implementing solutions to practical problems usingmachine learning and optimizing algorithms. Most recently, automated personnelscheduling subject to regulations while eliminating 80% of employee dead timewith respect to business needs.#Java, #Drools, #OptaPlanner, #JOptimizer, #Guava, #Git, #Algorithms, #Heuristics, #MathematicalProgramming, #OperationsResearch, #WorkforceManagement

Migrated and maintained operations. Redesigned and begun major redevelop-ment of site/app with some hand in design and business decisions. Part of thisjob involved merging a three-years old code split.#Frontend, #Backend, #Apps, #Agile, #Angular, #Bower, #NodeJS, #ExpressJS, #Loopback,#MongoDB, #SQL, #SEO, #UX, #BootstrapMaterialDesignAndSuch, #PHP, #Apache, #Git

Topic: hardness of approximation (Theoretical Computer Science).Primary contributions involved circumventing core obstacles or finding pivot points in proof techniques. Topics included approximation, average case, heuristics, and computational complexity.Select courses: Computational Game Theory, GPU Programming, Algorithmsfor Advanced Processing Architectures, Quantum Computing, Approximation Al-gorithms, Computational Number Theory.Taught: Program System Construction with .NET, Information Retrieval,Decision-Support Systems, Programming under Pressure, Advanced Algo-rithms, Object-Oriented Programming, Algorithm and Complexity, Introductionto Computer Science, and Algorithms, Data Structures, and Complexity.Advisors: Prof. Johan Håstad, Prof. Viggo Kann.

Simon's Institute program on Real Analysis in Computer Science. A massive and prolific gathering of researchers working in one field.

Elder care - chatting, cooking, cleaning, and other assistance. Summer job.

Summers '03, '04, '06.