Bryce is a Data Scientist and ML/AI Engineer who accelerates innovation through data design and integration in delivering solutions for clients. With 5+ years of experience in Data Science, Data Engineering, Machine Learning, Data Analytics, Python, and GCP, he is passionate about creating innovative and impactful technological solutions for businesses in various sectors such as fintech and eCommerce. As a data expert, Bryce oversees production implementation and validation of data science and machine learning solutions using a combination of techniques and methods to drive deployments. He builds, trains, and draws inferences and handles the release of pipelines within the tenets of industry standards on projects.
Delivering data warehouse and ETL solutions as part of an agile team using advanced ML techniques to improve performance and processes.
Helping build and improve infrastructure, application and performance development, and ensuring tight security including data encryption, security groups, and environment scanning.
Ensuring high-quality deliverables and implementing DevOps and security best practices in fast-paced environments.
Engineered cutting-edge AI solutions to optimize student success, designing and implementing an advanced Generative AI-powered agentic system that could detect students' subject interests, autonomously select promising topics, and deep dive into relevant concepts.
Developed a sophisticated agent architecture that generated personalized learning paths along with custom educational content and assessment quizzes to validate student understanding.
Architected a robust agentic back end leveraging LangChain, LangGraph, and BAML, with data persistence handled through Supabase integration.
Built an intuitive user interface using Streamlit for students' seamless access to a personalized learning experience.
Developed RESTful API endpoints using FastAPI to enable communication between the agentic back-end and front-end components.
Collaborated with educational experts to refine the system's ability to align AI-generated content with pedagogical best practices.
Managed the whole infrastructure on a dedicated server using Docker Compose.
Led Data Science initiatives to enhance the student conversion lifecycle, conducting in-depth analysis of the student conversion funnel, identifying key optimization opportunities, and implementing data-driven solutions.
Developed a sophisticated survival analysis model to predict long-term conversion patterns, accounting for the extended decision-making journey of prospective students.
Collaborated with cross-functional teams to integrate analytical insights into strategic decision-making processes.
Delivered actionable recommendations to head of data and CTO that improved student engagement metrics and optimized marketing resource allocation.
Worked on an Airflow data engineering pipeline to compute predictions on a daily basis.
Led the Data Engineering team to build and maintain a modern data infrastructure and to help business owners make data-driven decisions.
Developed custom data connectors and designed layered BI transformations to facilitate data usage.
Coordinated efforts to build a modern data infrastructure, including an ELT datalake and data warehouse, and acted as primary support for the Data Science team.
Oversaw projects, managed resources, and communicated with stakeholders to ensure the success of the overall mission.
Assisted the Data Science team in deploying models and resolving technical issues.
Helped the Data Analytics team in building analytics models, crafting advanced analytics, and ensuring maintenance of the data infrastructure.
Contributed to a notable improvement in data quality and providing effective tools for business owners.
Helped recruit new team members through a rigorous selection process and upskill the current team members to strengthen the group's skills and expertise.
Developed predictive time series models using LightGBM, validated and adjusted these models to ensure their accuracy and relevance, and integrated them into marketing processes to improve overall performance.
Rewrote the entire feature calculation algorithm, allowing for scalability and calculating over 500 features on a volume of 1 billion rows.
Improved the Google Shopping algorithm by evaluating 2M+ products daily to adjust the maximum budget according to the devices used, boosting growth and customer acquisition and making a considerable impact on ManoMano's growth.
Conducted an in-depth analysis of delivery time estimates at ManoMano, discovering that they were too pessimistic compared to actual delivery times, and recommended tracking data for delivery time calculations, contributing to a 3% increase in the overall conversion rate.
Built a pipeline to attract traffic to the ManoMano website, handling API connection, keywords retrieval, and bidding model.
Implemented new features and enhanced the Google Shopping SEA ML pipeline on the DIY platform.
Produced and implemented solutions to migrate ManoMano's platform to a new Airflow data engineering system.
Measured the actual effectiveness of models using A/B tests and collaborated with business teams to ensure project success.
Maintained the Data Science team's infrastructure, helped push the team's models into production, provided ongoing support to business teams on their challenges, and participated in Data Science competitions.
An AI Agent chatbot system powered by OpenAI that can detect students' interests, deep-dive into one of them, and generate a complete personalized learning path with content and quiz to validate student understanding. The goal is to gamify learning to make it easy for students to learn and be successful. The project was built upon a framework that makes it easy to create any AI agent for the school, the framework itself is built on top of LangGraph and LangChain, two industry standard library to build SOTA AI Agent.
Developed the Google Keyword Bidding model on the ManoMano platform, creating data pipelines to connect the Google AdWords API and establishing a list of keywords relative to the business and a bidding algorithm to drive marketing campaigns. Improved the solution efficiency from 10K a month with 0.97 profitability to 10K a day with 0.90 profitability.
Worked on the complete rewrite of a 5-year-old algorithm, delivering a scalable algorithm capable of computing features on the project. Added new features that led to a 20% uplift in the Machine Learning metric on the algorithm.
Handled full-stack solutions for NowLedge, a React.js-based SPA that allows users to upload digital documents (PDF) and take notes efficiently on the go. The app displays highlights to users on a nice wall with highlighted text and user notes in a bullet-point format.
Education
MSc (Hons.) Computer Science and Software Engineering