Senior ML Engineer - Customer Support· Helped to establish and lead a new team responsible for curating customer support data and building applications leveraging ML to improve the CS experience.· Designed and built a proactive solution for monitoring public comments for customer support issues, leading to reduced effort from support staff and increased customer satisfaction.· Led an internal effort for routing customer support cases and identifying emerging support case trends.Senior ML Engineer - User Behavior Analytics· Designed and built a causal inference system to discover user behavior patterns that lead to retention and engagement. Designed interventional studies to drive these behavior patterns.

Online user behavior analysis and prediction

Playing the role of CTO, Full Stack Developer, Lead Engineer, I'm building a platform for connecting passionate travelers to communities centered on sustainability and spirituality around the world. Key technologies: Node.js, Angular, MongoDB, AWS.

At Preact I invented a number of algorithms for customer behavior analysis and prediction from raw click events on SASS applications. For instance to assess the likelihood of abandonment, I built a MAP probabilistic model and to cluster a subsequence of events into semantic actions I used Non-negative matrix factorization.

Researching information retrieval, topic modeling, sentiment analysis, semantic analysis, and collaborative filtering for a developing news and entertainment portal.

Research and design of computational methods for inferring the three-dimensional coordinates of genes within in-vitro cells. Extensive use of python, bash scripts, and sun grid engine (cluster management) for massive scale computations.

Invented some low-dimensional regression techniques and applied them to color science problems. Wrote a thesis about it.Designed a data repository and website for similarity‐based machine learning and worked with a web development expert to implement it in ASP.NET.Transferred my research in fast low‐dimensional regression to a cross‐platform REALbasic application for the company Chromix. This included porting code for the solution to general quadratic programming problems.Designed a bash script library for distributing and collecting similarity learning simulations on a cluster of >400 nodes managed by Torque/Maui.

Received the highest teacher rating for this course to date and was nominated for the “2008 Outstanding Teaching Award” in the Electrical Engineering Dept.Created a set of interactive lecture slides used by subsequent instructors as well as a secure/interactive online gradebook using PHP and the Google Docs API.

Developed a python library to interface with an OpenRAVE simulation of a robotic arm and hand outfitted with capacitive sensors in the fingertips. This allowed for accurate and fast comparisons of the predicted capacitive coupling between sensors situated in a non-trivial geometry with ground‐truth measurements.

Designed an electrostatic simulation toolbox in Matlab, formulating the capacitance between 3‐dimensional conductive objects as the solution to a Linear Program.

Developed an “Enhanced Presence Information” Instant Messenger plug‐in, routing activity data inferred from a Bluetooth sensor board to a user’s contacts through the XMPP protocol.