UC San Diego

Columbia University

Carnegie Mellon University

Principal Data Scientist at Nike

Nike

Principal Data Scientist

Mirador was acquired by CUNA Mutual Group in December 2018.  

I continued in my role at Mirador as part of CUNA Mutual Group.

CUNA Mutual Group

CTO, Mirador

Mirador was acquired in 2018 and is now a wholly owned subsidiary of CUNA Mutual Group.

Mirador engineers cutting-edge technology to help small business lenders make faster, smarter and more profitable loan decisions. Backed by proprietary, machine-learning algorithms that provide real-time, continuous analysis of thousands of borrower statistics, our systems enable lenders to significantly reduce costs and loss rates while lending more broadly.

Raised initial and subsequent funding rounds (Seed and Series A).  Lead company Machine Learning efforts in credit modeling, web/product optimization, network routing, and other product enhancements.  Directly manage Information Security and Security Operations (constructed in AWS), IT Operations and all supporting infrastructure, as well as direct our Audit efforts through multiple years (SOC1 type II and SOC2 type II).   Support sales efforts though client presentations, thought leadership, and strategy.   

Wear other hats to support....whatever is needed

Mirador Financial

Co-Founder & CTO

Design and implement machine learning algorithms for novel use of flight and operational data.  Leverage company’s unique position as aggregator of disparate sources of flight data to create a global model of flight network activity.   Integrated additional 3rd party data sources to expand dimensional coverage of the network model. (i.e. NOAA METAR, TAFs, SIGMETs and AIRMETs for weather phenomena; Flight Tracker data for equipment tracking and real-time flight updates, etc.);  Used network model as inputs to Machine Learning algorithms which predict flight delay and flight propagation delay from 48hours to up to 6 hours before flight departure.  

Research included model explainability and individual prediction explainability techniques.  Constructed distributed data processing pipeline in AWS including an Active Learning environment in Python, Java, scripting, data engineering, etc.

FlightStats

Data Scientist

Expanded classification, content filtering, and information retrieval techniques for a learning and personalization platform built on NoSQL (Cassandra) and Python.  Developed new methods in adaptive learning and relevance determination for web--scale content personalization.  Work with other members of the development team on overall platform performance, reliability, and scalability.

Trapit, Inc

Sr Machine Learning Engineer

Responsible for all implementation, technical planning, strategy, and process of ai-one's core neural net based product offering.  Managed remote teams in Switzerland and in Northern Africa.  Developed short- and long- term product roadmaps.  Collaborate with research staff to evaluate fundamental technical capabilities and to bridge core technology to product to address real customer need. Provide pre-sales engineering support in the form of demonstrations and one-on-one technical sessions with potential customers.

ai-one inc.

Director, Product Development

Principal researcher to analyze data from thousands of vehicle recorders capturing millions of event streams from audio, video, GPS, and accelerometer sendor data.  Apply machine learning across manual analysis pipeline.  Prototype implementations in Matlab and C#/.NET/C++
* Performed computer vision and object recognition using OpenCV and C++  on compressed VGA video signals which recorded both road exteriors and cabin interiors.  
* Build classifiers to find salient events from audio streams (speaking, glass breakage, vehicle roll over, etc.)
* Built detectors for recognizing driven vehicle dynamics from accelerometer data using a variety of signal processing features; classified high level driver context and behaviors.
* Designed active learning, co-training, and human in the loop processes to drive on-going system improvement across all signals (i.e. allow the system to learn from its mistakes)

DriveCam

Software Engineer - Advanced Research

Managed a multidisciplinary team of designers, developers, and program managers responsible for all user experience and frontend development of Datek's numerous retail brokerage products including realtime market visualization tools (Streamer product line), clearing systems, customer online applications, and active trader tools.

Datek Online

Vice President, Client Development

Developed custom workflow solutions for clients using open source tools. Directed all technical strategy, development and infrastructure for both our company and our clients.

1bigthink / BGR Enterprises

co-founder

Responsible for finding, evaluating, and integrating new technologies for use in prepress operations.  Managed development and integration across multiple locations.  Developed prototypes for customized content management systems.  Generated sales presentations and supported sales efforts.

American Color Corporation

Manager of Technology

Long term project assignment as lead designer and eventual development manager for the Federal Direct Student Loan Consolidation System (FDSCLS).  Responsible for analysis, system design, UI design, and implementation.  Managed development team of 5 programmers.
Conducted formal acceptance testing with government leads.

Deloitte & Touche

Sr. Consultant

Built a data driven publishing solution for retail advertisers.  Developed workflow tools to allow customers to automate their ad publishing process. Designed and programmed a kiosk prototype using publishing system as core.

William Beaver