Boston University

Actively seeking software engineering positions.

Stealth

Enterprise Data and Machine Learning team within Analytics & Behavior Change
· Individual contributor to Campaign Production Launchpad system using GCP Spanner, BigQuery, Python, Node.js and React.
· Individual contributor detecting and fixing vulnerabilities in GCP pipelines using Github workflow, Jenkins and Checkmarx.

CVS Health

Senior Cloud Engineer

• Lead developer for SQUIRREL: Software Quality Under Integrated Review, Revision, Evaluation, and Learning which is a GitHub Software Assistant powered by LLM.
• Lead developer for JIFF 2.0 - a typescript library for secure multiparty computation.
• Lead developer for Algorand wallet for Hello Tractor using Vue.js, Capacitor, and Kotlin. 
• Lead developer and architect for system redesign of WebMPC and Boston Women’s Workforce Council which implements Secure Multi-Party Computation protocols using Typescript, Jest, Cypress, Django, PyTest, Redis and Mongo.
• Individual contributor to authentication and notification microservices using Nest.js and Typescript. 
• Individual contributor to Climate Lies, created interactive graph visualization using Three.js.
• Individual contributor to development of service for ASL recognition using Uvicorn, Strawberry (Python framework for GraphQL) and Apollo Federation.
• Individual contributor to Market Scan, Google Cloud Dataflow ETL pipeline for medical records using Python, Google Drive API and Google BigQuery.
• CI/CD tools and workflow used across projects: Docker Compose, GitHub Actions, and Portainer. 
• Other duties include designing systems, reviewing code bases, supporting junior software engineers, requirement gathering, project and grant proposal ideation, and leading technical workshops.

Boston University Software & Application Innovation Lab

Senior Software Engineer

• Developed the concept of end-to-end custody (E2EC).
• Developed web extension interface for managing digital artifacts, storing and recovering private keys using secure multi-party computation with Typescript.
• Invented a novel peer-to-peer communication protocol for establishing TCP between mobile devices using Kotlin, Capacitor, and Typescript.
• Created a proof of concept system for simulating distributed storage engine, demonstrating expertise in Node.js and Express.js.

dat.cash

Founder

• Developed image segmentation model for object tracking in flight systems using PyTorch.
• Synthetic data generation pipeline for training image segmentation model using Python 3.
• Developed pipeline for multivariate time series models and anomaly detection for objects in video using Python 3.
• Developed Grafana-based Docker infrastructure for monitoring and analytics.

XOR Security

• Created a system for generating an arbitrary amount of machine instruction data using Binary Ninja API, Python 3 and Multiprocessing. Before we generated the data manually using the Binary Ninja GUI. My system generated more than 20 GB of data in a few hours.
• Trained a generative long short-term memory (LSTM) recurrent neural network on machine instructions to learn representations using PyTorch. I then applied the encoder to generate unique ids given an instruction. These ids were used later by my team to color code similar instructions.
• Developed a training pipeline for experiment and hyperparameter tracking and reloading using PyTorch, MongoDB and Matplotlib.
• Researched applications of graph convolutional neural networks and language models for improving representation learning and transfer learning.
• Investigated integration of Relational Dynamic Influence Diagram Language (RDDL) server into our reinforcement learning environment using C++.
• Founded and organized STR's Machine Intelligence Club, which met to discuss recent machine learning research papers.

Systems & Technology Research

Associate Researcher

• Researched and read 72 research papers on previous and related work on distributed training for deep neural networks and privacy preserving machine learning.
• Conceptualized a novel decentralized and asynchronous training algorithm we call Hypergraph Stochastic Gradient Descent.
• Developed the peer-to-peer distributed training protocol and framework called Secure Multi-party Learning using Websockets, Redis and PyTorch on the Massachusetts Green High Performance Computing Center. 
• Presented preliminary findings to the Honda Research Institute in Offenbach, Germany who generously funded our work.
• My code is available here at https://github.com/ch3njust1n/smpl

Research Assistant

• Co-founded the Machine Intelligence Community, Inc. 501(c)(3) non-profit for democratizing knowledge and fostering the future researchers and engineers in machine intelligence.
• Cultivated communities at MIT, Boston University, Harvard College, and SRM Institute of Science and Technology, which to date have hosted 83 paper discussions, 65 technical workshops, 10 Society and Artificial Intelligence discussions, 21 community gatherings and projects across all our communities with both undergraduate and graduate students.
• Facilitated cross-community collaboration and events.
• Managed growth strategy and strategic partnership with Tech Together and industry sponsors.
• Led organizing committee for two annual student-focused conferences on machine learning and artificial intelligence, which attracted over 800 applicants from across 86 schools. Our first conference was held at the MIT Media Lab and our second at Boston University CILSE.
• Planned year-long conference strategy, outreach and weekly meetings.
• Organized conference keynotes and coordinated invited speaker travel.
• Designed conference badges, shirts, ads, and conference schedules.
• Managed conference finances and helped raise $63,000.
• Coordinated conference growth strategy.
• Managed corporate relations with Intel, IBM, Uber AI, G-Research, and Malong Technologies.
• Developed conference website and knowledge graph platform using VENoM stack, Vue.js, Express.js, Node.js, and MongoDB.

Machine Intelligence Community, Inc.

Co-founder and Co-director

• Implemented custom scraping tool to collect text data from the British Parliament website using Beautiful Soup and Python.
• Researched and reported on sentiment analysis and natural language processing solutions to understand correlations between political views and literacy.

Software Engineer 

• Developed Git commit section of data collection pipeline for our team's Mining User Software Enclaves (MUSE) DARPA program.
• Collected a 1.5 GB data set of 3,306,239 samples of Git commit messages.
• Designed and developed a data processing system for handling big data to generate a supervised data set for binary classification of bugs versus features using ElasticSearch, Multiprocessing and Python 2. 
• Trained a 1D convolutional neural network (CNN) to identify bugs in commit messages to ~83% accuracy using Torch7. 
• Applied trained network to generate a labeled data set for automatic source code repair.

Draper

Graduate Summer Intern

• Translated tests written by the Quality Assurance team into automation code in Java and Selenium.
• Maintained testing queues, triaged and investigated automation failures, and reported bugs for the development team using Jira. 
• Collaborated with the Quality Assurance team where I ad hoc tested and performed manual tests developed by the quality assurance engineers to find bugs for the Hitachi Content Platform Anywhere Android mobile application.

Hitachi Data Systems

Junior Software Engineer

• Meet weekly with Professor Dino Christenson on the Amicus Curiae Networks Project via the Undergraduate Research Opportunity Program. 
• Developed code to remove miscellaneous characters and standardize the data format using Java.
• More details about the project can be found at: https://amicinetworks.com/

Undergraduate Research Assistant

• Diagrammed existing code base for refactor suggestions.
• Transitioned QA build and test system from threads to heartbeats using Java.

Justin Chen