In October 2023, I transitioned from the HR solution team at PandoLogic (a subsidiary of Veritone) to the Veritone Advanced Research Project. In this exciting new team, I am primarily focused on the research and development of generative AI technologies. My core mission is to devise innovative solutions that can be integrated across Veritone's diverse business units, thereby enhancing our offerings and driving technological advancement within the company.

My journey with the Wade and Wendy team is entering a truly exciting, new chapter. We will keep doing what we have been doing — bringing Conversational AI to make the recruitment process intimate and pleasant — but with 1000x scale as part of PandoLogic! 🚀🚀🚀

At the end of 2018, I left Holler and took a new role at Wade and Wendy AI to focus on machine learning and NLP, especially in the conversational context. At Wade and Wendy, we built AI recruiters (chat bots) that conduct screening job interviews, gather responses and evaluate the overall fit of the candidates for better recruiter/candidate experience and higher efficiency. My responsibilities included:- Implementing a job description parser to identify different segments using text classification - Building machine learning models for NLP tasks for chat generation- Developing and/or maintaining micro-services to serve NLP features internally- Experimenting with transformer models, comparing the performance with classic models (e.g. SVM, CRF) for existing tasks- Designing and implementing large scale Elasticsearch datastore of people data for fast search

I started working at Holler as the second employee of the company. Being part of the very small team, I played various roles until we built our foundation and I could start focusing on data science. My responsibilities over the five years included:- Implementing a content recommendation system using clustering and ranking - Designing and implementing the data pipeline for analyzing user engagement and demographics- Leading the front-end implementation of the analytic application- Implementing the server-side API for the analytic application- Leading mobile (iOS/Android) SDK development- Implementing the server-side API for data exchange with our mobile applications

PsyQic was an iOS app where people shared predictions about various topics. I was the founder and the lead developer of the company, working with two other co-founders.

Negative Split Technologies provides solutions to marathoners and marathon spectators. The main product, ChaseMyRacer is a mobile (iOS) app to provide marathon spectators with essential information such as the estimated location of a marathoner and other important race-related locations.

I designed, proposed, and taught a new course, IST488/688 Social Web Technologies, where students learned how to design and implement social web systems using Ruby on Rails and various APIs. I also organized a programming competition, called SU App Challenge (http://appchallenge.syr.edu), where SU student developers demonstrated their app projects and competed for prizes (total $2,500). The event showcased innovations by some of the most talented SU student developers and catapulted many of them into the tech startup scene. I ran the event for four years while being a PhD student.

The Northeast Collegiate Triathlon Conference (NECTC) is a student-run non-profit organization that provides an organized series of triathlon races and for collegiate students in the northeast region. Founded in 2006, by three students from three universities across the region, the NECTC showed a strong growth over the years and involved over 15 schools and 200 athletes in 2009.I was one of the co-founders and the director of the organization until 2009.

My responsibilities included designing and programming for statistical analysis on transactional data from government agencies, more specifically: - Processing federal case record data, received from the Department of Justice - Managing metadata codes for the case records- Processing income statistics from the IRS - Processing population data from the Census Bureau - Generate monthly reports on the statistics of the data

- SELinux System Administrator Database: Analysis of the security-enhanced operating system using formal methods - Interface Based Intrusion Detection: Design and implementation of Linux kernel modules that simulate rule-based packet filtering on a network interface