As the Studio Project Lead at CMU, I've honed my expertise in machine learning and product management, leading a team to develop predictive systems for a major grocery retailer. Our collaboration, leveraging MLOps and technical skills, notably improved decision-making efficiency.
Led a team of 5 students in an out-of-shelf prediction system development for a large grocery retail client with a network of 250+ stores across multiple states.
•
Facilitated stakeholders’ communication among five global teams. Reduced communication delays and drove alignment by implementing regular virtual meetings; increased decision-making efficiency by 300%.
Machine Learning Product Management
•
Conducted four rounds of stakeholder interviews to uncover retail operational pain points. Derived specific ML system requirements to meet business objectives.
•
Analyzed client’s financial data; Built a compelling business case to persuade executives of the feasibility of an ML solution.
•
Coordinated machine learning engineering efforts. Prepared data for model input, evaluated algorithms’ functional suitability, and constructed MLOps pipelines on Azure; ML solution captured 30% of opportunity, generating more than 80% ROI.
Defined and documented product requirements for code review and repository organization settings features for a distributed version control product. The product supported over 20,000 developers at ByteDance after launching.
•
Executed agile methodologies to drive efficient project management, including sprint planning, backlog grooming, and daily stand-ups; facilitated cross-functional team collaboration and aligned business objectives with product specifications.
•
Conducted competitive analysis on the Chinese enterprise DevOps market, identifying and analyzing product gaps from six competitors. Provided insights to inform strategic planning for distributed version control Product-as-a-Service roadmap.
Designed and built a staging environment for all of Minerva’s portfolio products; streamlined deployment and quality assurance workflows and decreased production code bugs by 80%.
•
Spearheaded adoption of infrastructure-as-code practices to deploy cloud-native microservices; shortened development operations-related maintenance time needed by 70%.
•
Deployed and integrated credential storage vault into the company’s tech stack; eliminated 95% of hard-coded credentials and enabled intuitive credential access and storage for internal personnel.
Product Design and Development
•
Interviewed internal customers to articulate product requirements for a platform to streamline newsletter management flow. The product managed 100% first-party newsletters and over 615,000 subscribers and generated 20% of revenue at Minerva.
•
Identified data analysts’ pain points in existing SQL workflow, then designed and implemented a new workflow management product using Airflow; increased data analysts’ SQL error detection and response efficiency by 50%.
