Stratyfy is a transformational company that strives to empower decision-makers with performant and transparent products to build understandable, resilient ML systems and enable lenders with the power of data-driven AI/ML solutions to build a more equitable finance system.

➢ Led an organization of >100 data scientists and partnered with machine learning engineers and AI infrastructure experts to create swift, automated approaches to underwriting and operating credit card business while providing excellent customer service.➢ Created solutions to accurately identify, predict, and prevent credit card fraud via machine learning and artificial intelligence. Prevented revenue losses from criminal activity and generated major value out of an underutilized data set. ➢ Achieved significant minimization of the loss-to-sales ratio, significantly bolstering operating efficiency due to efficacious algorithms produced by my team.➢ Created major commercial value from underutilized credit card utilization data (inclusive of CapitalOne-specific and external data sources), significantly raising revenues from data-driven underwriting (deciding whom to extend credit to) and credit-line increases or decreases.➢ Spearheaded conversational AI development for the credit card business, setting strategy for the development of compliant, efficient development and deployment of transformer-based large language models.➢ Built an organization premised on equity and inclusion, while maintaining a robust, upbeat culture that cultivated excellence. Right-sized organization to maximize job-satisfaction, productivity, and quality.

➢ Led a team of eight data engineers and four data scientists, partnering with product, hardware and firmware engineers to build a scalable platform for data-driven product development, devising AI-systems that accurately predicted customers’ grocery needs via proprietary technologies.➢ Oversaw the development of a robust conversational AI platform, culminating in a winning customer experience that resembled interactions with a live human concierge.➢ Built and managed a high-performing team and absorbed another organization’s engineers despite low morale during major reorganization, establishing a team known for producing the most robust and high-quality data products in the company.

➢ Developed proprietary, personalized, and lucrative auto-replenishment system for home grocery deliveries.➢ Created a novel method for generating image-based product recognition datasets. Co-invented pressure-based IoT device to identify and track usage of grocery items.

➢ Managed a cross-functional team of engineers and data scientists that created lucrative new approaches to client acquisition and retention for private bank.➢ Led the development of algorithms that accurately identify strategic new geographies for the bank to acquire business from high-networth demographics.➢ Oversaw a cross-disciplinary team to create predictive models that identified and retained new customers for wealth management, ushering profitable new approaches for client acquisition.➢ Led the team that built SDLC automation for business users, enabling business to self-serve their business intelligence and analytic needs.➢ Managed the production of a vendor-independent data visualizations library that is compatible with Qlik and other vendor tools.

➢ As part of a program to enable academics to transition to industry, built a web-based tool deployed on AWS for nonprofit organization GuideStar to analyze financial viability of nonprofit organizations using logistic and survival regression.

➢ Built a MATLAB-based analysis package for 120 frame/second megapixel Charge Coupled Devices (CCDs) for X-ray detection for very fast X-Ray and electron microscopy. Developed software to take a scrambled binary file of 16-bit pixels, recreate the image, calibrate the detector data and analyze the result. Developed and maintained the Qt-based user interface and linux systems on which the data acquisition system runs.➢ Completed a proof-of-principle experiment, including writing a software-based multi-channel analyzer in MATLAB, to develop a new technique for microphonics noise cancellation in radiation detectors for nuclear science experiments. Used a digital algorithm on an FPGA to measure the microphonics noise on a detector and create an adaptive filter in real-time.

➢ In a collaboration of over 5,100 scientists, one of five technical experts responsible for the hardware and software functionality of a sixty-two million readout channel particle detector, contributing to the discovery of the Higgs boson.➢ Developed a novel method for identifying the absence of strong gravity (black hole) events in the earliest Large Hadron Collider data from the ATLAS detector.

➢ Taught undergraduates in small group tutorials, graded papers, and tutored students in Part Ib Quantum Mechanics, Part Ib Experimental Methods, and Part II Particle and Nuclear Physics.

➢ Worked on research and development of a new particle accelerator, the Small Isochronous Ring (SIR) to study the longitudinal-radial space charge effects that plague high intensity synchrotrons in a small, low-energy experiment.➢ Measured lifetimes of rare nuclei far from stability using a segmented germanium array (SeGA) detector, making ground-breaking measurements in the doubly-magic 100Sn region.

➢ Taught, graded, and supervised Physics 191: Physics Laboratory for Scientists.➢ Taught Integrative Studies in Physics 209: Mysteries of the Physical World.

➢ Developed an analysis procedure for a special kind of exotic physics decay in C++, ROOT, and the ATLAS software package Athena, implemented on both the local computing clusters and on the world-wide ATLAS Grid during its initial testing period.

➢ Used Monte Carlo and Geant4 simulations to model neutrino interactions in the MINOS detector, and ROOT-based C/C++ code to separate noise and background from the signal and measure the rate of rare interactions.