CTO and Co-founder
2018 - PRESENTADEx- Selected for the 2019 class of Colliers Proptech Accelerator powered by Techstars.
- Built and led a team of up to ten engineers to build an AI platform for automatically summarizing and reviewing users' documents.
- Developed technology partnerships with companies like Box and Salesforce.
- Obtained and maintained giant enterprise clients in the real estate space.
Technologies: Elasticsearch, Spring, Play, Neo4j, MongoDB, Jira, AWS, Google Cloud Platform (GCP), Docker, Flask, Python, Deep Learning, AngularSenior Software Engineer
2016 - 2018LinkedIn- Taught learning-to-rank methodology to other engineers, as one of LinkedIn's AI Academy's inaugural teachers.
- Drove the targeting of two-way interest for search results (i.e. candidates accepting inmails), instead of just recruiter engagement (i.e. recruiters sending inmails). This caused fewer complaints from users bombarded by irrelevant inmails.
- Patented several techniques including document parsing, factorization machines, and impression discounting.
- Migrated our team's offline training dataset generation pipeline from Pig to Spark, to allow for easier testing, debugging, and maintainability.
- Created a model which incorporated impression discounting features for the Recruiter Search product. This caused more inmail accepts and a healthier overall Recruiter Search ecosystem.
Technologies: Java, Machine Learning, Apache Pig, Spark, Scala, Deep Learning, Data Analysis, SearchSoftware Engineer
2015 - 2016Connectifier- Was integral in its team of ten developers, allowing the company to be acqui-hired by LinkedIn for over $100 million.
- Improved Connectifier's company canonicalization as well as retrieval of data from crawled resumes.
- Added to the company's AutoSearch extension a way to show Connectifier search results, even upon a user's searches on other sites.
Technologies: Java, Play, Angular, Scala, Machine Learning, Factorization Machines, MongoDBSoftware Development Engineer II
2013 - 2015Amazon.com- Rolled out machine learning models that improved Amazon's forecasts of demand for products under promotions, for all world regions.
- Integrated IPyNotebook running on PySpark with our platform so that software developers, research scientists, and data analysts could access, transform, and share data effectively.
- Designed and deployed software that allowed downstream clients to bias-correct our team's forecasts in bulk.
Technologies: Java, Hadoop, Spark, IPython Notebook, AWS, Random Forests, Machine Learning, Software Engineering, Apache Hive