Patrick Philips Email & Phone Number

San Francisco, California, Us

San Francisco, California, Us

Head of data, responsible for data science, data platform engineering and data analytics. Senior leader reporting to the CPO and working closely with other executives to define a strategy and roadmap to use data to tackle a huge world problem: how do we take better care of our parents as they age?Honor is the home-care solution that helps older adults continue to live and age in their own homes with greater joy, comfort and grace. Funded by Silicon Valley trailblazer Marc Andreessen, Honor combines the best care professionals in the industry with technology that modernizes the entire home-care experience, making it more personalized, flexible and accessible for family caregivers and their aging loved ones.

San Francisco, Ca, Us

Head of data, reporting to the CTO and responsible for building and leading a diverse team of data scientists, epidemiologists, data engineers and analysts who are driven by our mission of tracking and curbing the spread of contagious disease.We use smart thermometers and a mobile app to create a sensor network providing real-time, geolocated, clinically accurate health data. This network and the predictive algorithms developed by my team allowed us to track unusual patterns of febrile illness in NYC and Florida during the early pandemic when widespread testing was unavailable. Our work was featured on the front page of the New York Times and contributed to a step change in the business as measured by booked revenue.

San Francisco, Ca, Us

Established a Data Science team using billions of dollars of historical spending to build products that help our customers use the cloud better.Highlights include:Standardization - Parse useful/relevant attributes out of TBs of raw spending data so that humans can make sense of their billing data. Currently powering our core platform analytics engine.Anomaly Detection - Scalable, performant system to identify unusual spending patterns and push alerts directly to customers with minimal latency (AWS Fargate)Rightsizing - Generate recommendations for resizing resources in AWS and Azure based on historical usage patterns to optimize for savings and risk (AWS EMR, Spark, Spot instances).Reserved Capacity - Generate recommendations for reserved capacity purchases based on historical usage patterns to optimize savings (AWS Lambda)Forecasting - Developed algorithms to predict future spend across AWS, Azure and GCP "in a single pane of glass". Shipped a feature offering high precision estimates for the current month as well as an interactive tool to forecast many months into the future with minimal latency, and taking into account cohort growth patterns identified from billions of dollars of historical spending data.Workload Placement - Help customers to model migrations costs when moving from on-prem to cloud as well as across major public cloud vendors (AWS, Azure & GCP). In-house benchmarking to compare performance across current and legacy hardware.TrueCost Explorer - Simple point-and-click interface to help executives and non-technical users easily browse their spending data and answer questions without needing to understand underlying complexity of billing file schema across multiple vendors.Data Warehouse - Rearchitecting our data warehouse using Amazon Athena has decreased our costs by a factor of 50 with no impact on performance, while adoption has increased significantly.

San Francisco, Ca, Us

Leading a team of data scientists focused on modeling consumer behavior from passive wifi observations to generate insights for brick and mortar retailers. We're using Scala and Spark for our production pipeline, Mesos and Chronos to manage our EC2 clusters and an unholy alliance of Kafka, Cassandra, RDS, S3 and Redshift for data storage.As an individual contributor:Implemented a revised Spark algorithm for our production classifiers, achieving comparable performance while reducing training time from double-digit hours to a few minutesLed a project to overhaul our Scala production pipeline, reducing the runtime for the nightly batch jobs from 10+ hours to under 4 hours. This project also reduced overall AWS costs by 45 percent and reduced number of pipeline incidents by a factor of 10.Developed a real-time pipeline to measure line wait times using Kafka, Spark Streaming and Cassandra. Achieved over 95% accuracy based on a validation set collected using TaskRabbits and a custom data collection app.Worked with front-end & product teams to design, test and implement customer-facing training data collection tools.Built custom data dashboard using python, matplotlib, pandas and flask to give non-technical colleagues the ability to schedule analysis, view curated data visualizations and identify common data issues.

Sunnyvale, Ca, Us

Part of the Product Data Science Organization under Daniel Tunkelang. The group included a number of leading Data Scientists, including Monica Rogati, Joseph Adler, Gloria Lau, Vitaly Gordon and Peter Skomoroch.As the crowdsourcing lead, I led the efforts to develop a self-service crowdsourcing solution for data scientists, software engineers and product managers. Specific projects that we improved using human evaluation include: + developing a transition based model for title and seniority standardization+ disambiguation of LinkedIn members in new articles ("People in the News")+ multi-lingual comment spam detection+ personalized search relevance evaluation+ free-text topic identification for LinkedIn Pulse+ skills identification/standardization for endorsementsAdditionally, I developed models to evaluate/predict crowdsourcing worker performance based on a variety of historical and performance features, in addition to existing industry-standard models. I also led the efforts to partner with a leading crowdsourcing platform provider rather than building one in-house, identifying leading crowdsourcing providers, evaluating vendors, shepherding a multi-year contract through procurement and legal, and managing the vendor relationship.

San Francisco, Ca, Us

Early employee at CrowdFlower, focusing on client relationships, product management and sales engineering. Favorite projects include working with engineering to develop "DoublePass," a fundamental reworking of how we manage quality for subjective crowdsourcing tasks. I also managed three of the five largest accounts by total revenue.Later in my time, I focused on evangelizing crowdsourcing to non-technical business executives and senior decision makers, including delivering presentations/demonstrations, developing proof-of-concept applications and managing the full sales/marketing cycle.

Eugene, Oregon, Us

One of two members of the first on-the-ground team in Uganda for a non-profit providing business skills training to female entrepreneurs. Key achievements include identifying target community to locate organization, building partnerships with first client groups, adapting training curriculum to meet local needs and negotiating lease for organizational headquarters.

Cambridge, Ma, Us

Financial consultant for civil and criminal environmental litigation, providing general litigation support, financial modeling and developing expert witness testimony. Focus areas include Ability-to-Pay determinations and Owner/Operator Liability under CERCLA, Clean Water Act violation penalties and fraudulent conveyances under the Bankruptcy Code.

Ba, International Politics And Economics; Mathematics (Minor)

Doctor Of Patience (Honoris Causa)

Political Economy Of Financial Crises