Phil is a UX designer, researcher, and product owner with ten years of experience. His work has led to an acquisition, an IPO, and a top app in the Play Store (2+ million users). Phil's specialty is solving complex product-market fit problems for enterprises—building business products people actually use. From ideation to user testing, prototyping, and the iterative development process, Phil loves the craft and working on teams. Phil brings a detail-oriented approach and an open mind.
Annoyed with poor auto-correction on my Android and interested in building a true ML project, my development partner and I designed and built an NLP keyboard that uses full sentence context to predict spelling and word correction.
Completed ML classes: "Machine Learning" and "Structuring Machine Learning Projects" by Andrew Ng of Stanford, as well as "Natural Language Processing in TensorFlow" by Laurence Moroney of Google.
Studied the latest research on corrective keyboards: Nlpprogress.com/english/language_modeling.html.
Designed a novel model using Peter Norvig’s spell-checker and Google's Bidirectional Encoder Representations from Transformers. Our model used character prediction and full sentence grammatical context to predict and correct words.
Leveraged open data sources then curated a custom data set using Enron emails and Twitter dataset to cover both the professional and casual typing use cases.
Built a program to run our full sentence model against our data set and saw improvements over the current Google NLP Keyboard. But our processing time was too high for a mobile device. It was fun. If there was a business here, I would have pursued it.
Technologies: Deep Neural Networks, TensorFlow, Natural Language Processing (NLP), Artificial Intelligence (AI), Machine Learning
Designer | Founder
2015 - 2020
Founded Sesame which was named "a must-have app" by TechRadar and touted by Android Unfiltered to "change how you use your phone." It has 2+ million installs, 25% long-term retention, and is a top revenue-generating productivity app.
Conceived and designed a data service that connects APIs and native sources to power a universal search for Android that is filled with deep links into your apps (Contacts, WhatsApp, Spotify, Slack, Reddit, Instagram, Netflix, and dozens more).
Grew our user base from 0 to millions by hyper engaging a niche of power users. Followed the lessons from "Crossing the Chasm," built a community of 30,000 beta test users, pushed weekly releases, and iterated our way to a good product.
Developed a coded search algorithm that lets you navigate to pretty much anything in one or two taps. Leveraging device usage data and matching the first letters of words "S" and "N" will pull up "Spotify: Nevermind." Next time just "S" will.
Managed a remote team in Taipei and India and organized and managed the project.
Designed an onboarding UX that took us from 3% to 27% long-term retention. This was no easy task. The key was finding the 'aha' moment that converted power users. We did this through relentless experimentation and listening to user feedback.
Led business development efforts to partner with Nova Launcher, resulting in Sesame being natively distributed by an app with 10 to 50 million users.
Deeply integrated it with the Android OS. I have an intimate understanding of what's possible with Android.
Was offered $1+ million acquisition by two well respected companies.
Founded Suited, a tool to improve hiring outcomes by correlating interview answers to job performance. It's like a SaaS version of Google's internal hiring tools: Bit.ly/Suited_Vision.
Started with science. The problem with interviewing is people ask different candidates different questions and don't use a consistent grading methodology. Ran concept validation tests on a solution and heard "I want this" from HR managers.
Recruited a technical co-founder and two organizational psychology PhDs with 25+ years of experience each.
Created a robust library of competencies and interview questions proven to predict performance for dozens of jobs. Seeded using the web service from Onetonline.org, curated it with Org Psych PhDs, and turned it into a company editable wiki.
Inspired by popular tools like Gong.io, we designed an automated transcription service to take coded notes and point to evidence in your grading decision. We removed the tedious steps and output a proper hiring packet for each candidate.
Signed two pilot customers: Nvidia and Krimsey's Cajun (restaurant chain).
Partnered with Venrock, Luv Ventures, First Round Capital, and Sequoia to explore bringing evidence-based hiring to their portfolio companies.
Technologies: Web, InVision, Sketch
2017 - 2017
Started as a freelance designer and was offered the co-founder role after two weeks.
Led user testing, UX, UI, visual design, and roadmap management.
Developed ClozeLoop, a product that layers Quora style curation tools onto Google Docs. Coworkers can mark docs helpful, tag them, and ask and answer questions in threads. It’s delivered through Slack.
Initiated user test their existing product. I made sure to be unbiased so the tests were believable. Watching this, the founders themselves concluded a pivot was needed.
Ran Google-style design sprints with the team including engineers. The group thrived in this structured ideation exercise. We agreed to a single sentence value prop and designed a prototype to test it. I built it in InVision and we user tested it in 1-hour scripted sessions. With new confidence in our trajectory, we started coding.
Managed an 8-week customer pilot. Embedded with a 15 person sales team, we shipped a new version of ClozeLoop each week. Improvements can sometimes be slow but if you keep at it and listen to your customers, eventually your breakthrough. At the end of our pilot, the customer wrote us a 5-figure check to keep the software.
Built as a side project and designed to be as simple as possible. Ride Fair shows surge and ETA information for all nearby Uber and Lyft rides. One tap takes you directly to the ride of your choice. It queries the Uber and Lyft APIs in real time. There are some neat design tricks: the background is a stylized version of Google Maps, showing your real location.
Conceived, designed, and developed by me and Steve Blackwell. 4.7 rating in the Play Store.
Featured in Fortune (For.tn/2h1kyW7) and the SF Chronicle (Bit.ly/RFchron).
Technologies: Lyft, APIs, Google Maps, Android, iOS
UX/UI and Product Lead
2012 - 2015
Served as the lead product owner and UX designer for the core web product as the company scaled from five to 250 customers, millions of users, and exited via IPO. I was accountable for user acquisition, retention, and the overall success of the platform. My ultimate goal was a 0% customer churn. To accomplish this I focused on user-centered design, product/engineering efficiency, and testing quality (QA is the bedrock of any well-designed product).
Managed on-site and remote teams of more than seven people to coordinate complex projects across many departments (engineering, product marketing, implementation, customer support, and legal). I was responsible for translating business goals into software features and earning stakeholder trust in a company that grew to over 400 employees.
Introduced Agile development to the company and a disciplined build, measure, learn approach as detailed in the book "Running Lean." I started first with my team, within a year the entire company ran Agile and many ran iterative experiments.
Improved our registration flow conversion from 20% to 60%. We replaced the registration form, which required users' SSN, with a single sign-on button branded by their employer. We then embedded SSO registration inside their health benefits enrollment (instead of wasting effort to lure in users, we went to where they already were engaging). To do this we built an iframe and API to conduct the secure handoff.
Designed a first time experience that had users complete a search and see how Castlight saves you money. Before rushing into some complicated walkthrough experience, we looked at our analytics data to look back and see what our power users had done the first time they used Castlight. This showed us that conducting a search is the key behavior. We then used location, personal claims data, and analyzed their health plan options in order to present them with a simple flow that converted first-time visitors into repeat actives.
Designed a simple UX/UI for a powerful search feature that pulled data from multiple sources simultaneously. Shopping for healthcare is confusing. It's hard to know what a "good doctor" is. Quality of care is more important than cost when choosing a doctor. We leaned on UI conventions from comparison shopping sites like Kayak, then took time user testing what type of information helped users discern which providers were high quality.
Built a claims-based provider review system. When users saw a doctor, Castlight would prompt them to review the care they received via an email or in-app prompt. The reviewed survey was based on the CG CAHPS methodology which rating outcomes.
Partnered with Quest Diagnostics to embed appointment scheduling into Castlight. Users could see available times, input preferences, book, and receive a confirmation all from within Castlight. This required complex data exchange between both enterprise companies. API's had not been fully developed, we had to lay that groundwork.
Conceived and designed a care team that analyzes a user’s historical medical claims then recommends providers from their health plan with superior quality ratings, lower prices, and similar driving distance. It displays total savings opportunities and lets you easily access your family's providers along with benefits information.
Managed the "Castlight Guides" support team tools. This included a knowledge base and live posing tool that allowed them to securely screen share with the user. Designed and implemented a personalized email tool that used claims data to deliver relevant and timely emails. In some instances, we improved the click-through rate by over 1,000%.
Technologies: Microsoft Office, Plugins, Salesforce, Iframes, Single Sign-on (SSO), Content Management Systems (CMS), Web Applications, Google Analytics, Jira, Mailchimp, Machine Learning, Adobe Photoshop, Sketch
UX/UI Designer | Product Owner
2012 - 2012
Built custom integrations for AT&T, Best Buy, and Orange. Too many of their smartphones were being returned because people didn't know how to use them. DashConfig provided a first time experience that prompted users to connect to WiFi, sign in to their Google Accounts, backup contacts, etc. This early usage of their devices' most impressive features and reduced device return rates. We take mobile onboarding for granted today, but in 2012 these were cool ideas.
Managed an onsite and remote team development team.
Led ideation, design, roadmap planning, and execution.
Technologies: Analytics, Web Applications, Adobe Photoshop, Android
2010 - 2012
Let's Do This
Founded the company, recruited a team of six, designed the app, ran project management, and owned the analytics and marketing.
Shipped three products for the web and iPhone with some cool technology in them: a learning algorithm, group text functionality before GroupMe existed, Facebook integrated, and real-time features that tend to crash servers. Unfortunately, we tried to apply this technology to a convoluted UX of "what are your friends doing now?" It had a terrible chicken and egg problem and was dependent on users having a lot of motivation that just didn't exist.
Technologies: Apache, Messaging, Facebook Open Graph API, Web Applications, iOS