Doron Zehavi

I build high-performing wellness products with strong repeat purchase and a fast-growing subscription base. ANML started as a one-man operation in my kitchen, scaled into a commercial kitchen on the Lower East Side, and eventually transitioned to a contract manufacturer and 3PL once demand exceeded what I could produce by hand. We sell through Shopify DTC along with select retail and online partners. My focus is simple: create products that are clean, effective, and trusted enough that customers come back.

• Developed and launched a machine learning-based Anomaly Detector (Journal Insights) which has helped 100+ Workday Financials customers monitor millions of monthly Journal Lines in real-time in order to shorten their period-end close and mitigate financial risk in its first year. • Developed and launched a machine learning-based Anomaly Detector (Expense Protect) which has helped 50+ Workday Expenses customers identify and correct errors in employee submitted expenses in its first six months. • Initiated a strategy to pivot from supervised learning to unsupervised learning model which reduced time to value by 90% and increased adoption by 200%. • Initiated a strategy of using clusters and thresholds to show the most anomalous instances within a cluster which reduced total results by 98% in order to improve discovery and reduce user fatigue. • Developed an explicit feedback strategy to improve relevance of anomalies by boosting similar true positive and hiding similar false positives, as well as showing what feedback was submitted on similar lines. • Developed and presented a proposal to VP and Sr. Director of ML for an innovative Financials Search feature leading to a go-forward decision.

Team: ML for Financials • Developed and launched a machine learning Predictive Account Scoring solution used by Sales and Marketing to prioritize and assign accounts leading to an anticipated 15% increase in win rate and reduced time to close a deal. • Presented go-forward recommendations to end-of-life a legacy ML feature to key stakeholders.

Team: ML for Planning & Financials • Presented two breakout sessions, at Workday Rising 2018 in Las Vegas, to hundreds of customers on the topic of machine learning, the Workday Graph and the Skills Cloud. • Managed development and release of the Workday Graph which powers multiple new Workday applications. • Collaborated with the Talent application team to power the Skills Cloud application. • Sent out surveys and met with customers to vet multiple use cases for feasibility and business value. • Developed, documented and evangelized the new machine learning development process across the company. • Consolidated, organized and maintained the machine learning use case backlog.

Teams: Financial Projects and Adoption • Built robust Adoption Planning and Adoption Navigator features which were quickly adopted by 150+ customers as the second developer on the team. • Worked independently to set up the necessary infrastructure and initial development for a new product area. • Lead the company-wide refactoring effort for my product area resulting in 100% compliance.

Continuing research first started in my Senior Design Project, I am the lead developer of Virtual Front View, which is an Android application that streams video from one device's camera to another through the WiFi-Direct protocol. I modify the app to allow me to run tests and gather results on the reliability of WiFi-direct with multimedia data (such as streaming video). Additionally I participate in weekly research meetings where I share my results as well as present about published research papers that can be utilized to improve the reliability of the ad-hoc network.

Completed Courses: * BUS ADM X448 - 018 Managing Product and Go-to-Market Strategies