Advising a stealth team regarding their search productFreelance developmentLearning Unreal Engine (C++)Building a deduplicating backup tool. The files are decomposed into parts such that small changes to a larger file would only require the small change to be added (based on a technique from the rsync thesis). (Kotlin, LevelDB, SQLite)

ROSS Intelligence is a search engine for lawyers. The engineering team had 10-30 people during my time. Data pipelines. We needed a pipeline that could quickly update our indexes with 10M+ documents. Taught myself Apache Beam, BigQuery, and other big data tech. Created a pipeline for annotating documents (footnotes and citations) and storing results in S3 for production. Rebuilt most of the existing pipeline. Extracted statistics and made them available via BigQuery. (Apache Beam & Google Cloud Dataflow, BigQuery, Airflow, GCS, Parquet, Java, Python)Document analyzer v2 was part of an integration with Clio to increase awareness (2x usage) combined with technical strategy to adopt GraphQL. Created a generic async task service to coordinate queuing jobs and publishing results. Productionalized GraphQL, added authentication via JWT token verification, and enabled WebSockets via nginx. Improved the UX by using a library that replicates PDF style in HTML. (GraphQL, Redis & PubSub, SQS, Lambda, Python & FastAPI, Kotlin)Citations Ranker. SMEs felt that influential cases were not being ranked appropriately. I led development (with an intern) to build a ranking service based on citations. Created an evaluation data set from a publicly available data source that saved time and money. Wrote code to extract the required data and load into the production data store. (ElasticSearch, Python & Flask, CometML)Legislation search engine. When I joined the team, the indexing process was taking too long (> 1 day) and there were two sets of code. After listening and discussing with the team, we made it one set of code that could index within 2-3 hours. (Solr, Python, Grails, AWS Batch, S3)Refactored and added features to legacy code for self-serve registration that enabled 20% MoM growth in users. Mentored junior developers. (Grails, jOOQ, PostgreSQL)

Built system for fast speech recognition (iOS, Android, Python, PostgreSQL)Developed Android app for recording and watching videos (ExoPlayer, RxJava, Glide, SQLDelight, AutoValue, Retrofit)Designed search indices and optimized relevance (ElasticSearch, Java)Implemented a paper to extract professional skills from text (Python, Gensim, Java)Developed a “person vector” model with word vectors to find similar people (Python, Gensim)

Worked on a mobile game with 10M+ installs. Built Android app that has 1M+ installs.

Mentored team members to become team leadsLed teams building mobile apps for large clients and startups

- Technical lead for Android app with >500,000 daily users

- Developed imaging software that combines multiple images into a single panorama for use with a multi-lens camera- Developed with C/C++

- CPS615 - Theory of Parsing- CPS841/CP8309 - Foundations of Semantic Web Tech.- CPS109 - Computer Science I (Java Programming)- CPS125 - Computing for Engineers (C Programming)