Bryan Chen Email & Phone Number

Irvine, California, Us

Senior full stack engineer with experience in ML

Irvine, California, Us

- Incorporated correlation metrics from our open-source libraries into our internal UI product. Built custom support of these metrics into D3.js to allow for special behavior for UI charting- Performed unit testing of frontend code using Jest to ensure proper behavior with mock user interaction- Modified update callbacks to occur asynchronously in the backend with limiting requests, resulting in a 400% speed-up of calculation operations in the UI- Became a team resource for both UI code and OS/backend code for my team of 8- Integrated a new backend and frontend before a major product deadline, ensuring that API calls were properly made, saved, and later deployed in the project. Delivered product with limited resourcing and aggressive timelines- Mentoring a summer intern, guiding him through a full-scale OS project, and maintaining weekly discussions to develop tactical skills for the team and projects that aligns with career growth

Charlestown, Ma, Us

- Implemented deep learning models in Pytorch for image classification, detection, and segmentation on CXR images for Pneumothorax- Investigated the data requirements between supervised learning tasks and weakly supervised learning in order to determine the feasibility of using WSL for localization tasks - Trained and collected performance results for Pneumonia detection, using both supervised and weakly supervised models. Looked at the application of these models for COVID detection, fine-tuning on locally-collected data

Mclean, Va, Us

- Designed, implemented, tested, and documented a program to automate the conversion of non-standardized data sheets into a standard format- Achieved a 96.6% accuracy compared to the 81% accuracy of manual labeling, and increased speeds by over 24,000%- Created a solution that outperformed a third party in both scalability and speed, saving > $2MM for CO- Presented a product evaluation to the business team at the end of the summer to discuss feasibility, costs, and next steps

Hong Kong, -, Hk

- Created a ML model to automatically filter sparse and abbreviated emails into distinct categories, achieving a 98% accuracy on testing data. The model went into production immediately for testing and use.- Implemented a regex parser to extract data from emails and format the information into a clean output.- Developed and tested a separate ML model to extract information from tabular data in emails. Designed this model as a POC to illustrate that this task is achievable.

Cambridge, Ma, Us

- Created an automated comment moderator for MIT’s App Inventor to filter out profane and inappropriate posts.- Built a model for profanity detection using Facebook's FastText combined with NLP processing.- Edited open-source code to facilitate front and back-end communication to retrieve model results and display them on the site in real time.

Mclean, Va, Us

- Worked on creating efficient Parquet column retrievals from AWS S3. - Developed a POC and simple working model that highlights the potential of specific Parquet column requests from S3.- Parsed through large libraries of open-source code to determine the most effective method to gather the necessary information.

- Developed an Android app to simplify the creation of formations in dance, similar to a playbook.- Worked on creating an easy-to-use UI and UX design for the front-end, and simplified the planning portion of formations.- Linked the backend into FireBase in order to save and share formations with other users.

Mclean, Va, Us

- Attended in a one week long summit at Capital One where we learned crash courses in topics such as iOS and Android development, IOT, Git, and Machine Learning. - Participated in an all-day hackathon at the end of the week, where my team and I created a bill-splitting Android app using Capital One's Nessie API.

Boston, Ma, Us

- Deep-learning intern developing a new sentiment analysis system to function as an input for an AI, which will allow the AI to respond more accurately to the given situation. - Tested different model layers, such as BiLSTMs and CNNs, to determine the best model for the problem.- Achieved an 87% accuracy in sentiment classification for positive, neutral, and negative sentiments.

- Worked closely with counselors, lab assistants, and coordinators to introduce incoming freshmen to electrical engineering fundamentals. - Served as an advisor and helped freshmen transition from high school into the college atmosphere.

Kowloon, Hong Kong, Hk

- Researched the Braintree code to determine its compatibility and functionality with the Compathnion mobile app.- Worked closely with an advisor on website development to update the current online Compathnion website.- Provided the start-up with a usable website for them to integrate and expand upon.

Master Of Engineering - Meng, Computer Science

Bachelor’S Degree, 6-3: Computer Science

High School