Yumi He

I worked as a Software Engineer, collaborating closely with UI designers to implement the app screens they designed, focusing on realizing the front-end user interface and ensuring the product was visually cohesive and user-friendly.

• Explore Screen: Implemented the explore functionality that allows users to discover videos and photo posts based on their preferences. The explore feed prioritizes the latest posts and adapts to each user's browsing habits. Additionally, I implemented filters to explore posts from nearby users, school circles, or career circles.
• Events Tab: Built the events tab, which allows users to discover and join interesting events, both online and in-person. Users can join events with a simple button click, and event organizers can control the visibility of their events (e.g., restricting visibility to friends with level 5 or higher).
• Friendship Levels and Post Visibility: Helped develop a dynamic friend level system that starts at level 0 (not friends) and increases as users interact (follow, chat, like, comment, etc.). This system impacts the visibility of posts, events, and comment permissions, giving users more control over who can see their content.
• AI Chat Functionality: Contributed to integrating GPT-4O-powered AI chat features, allowing users to select an appearance and voice for their AI companion. The AI can engage in conversation, create quick diary notes, generate post titles, and suggest content, enhancing the user experience by providing convenience and a personal touch.

I worked as a Marketing Intern at lvlup.ai, focusing on increasing user engagement and addressing the issue of limited user adoption.

• Targeted Social Media Campaigns: Created targeted social media campaigns to attract different user demographics, such as students, professionals, and local communities, using relatable content and highlighting the app's unique features like event discovery and AI-assisted journaling.
• Influencer Partnerships: Partnered with micro-influencers in niche communities to promote the app, leveraging their audiences to increase brand awareness and credibility.
• Referral Program: Developed a referral program that rewarded users for inviting friends, encouraging organic growth through existing user networks.
• Community Challenges and Events: Organized community challenges and events to boost user engagement, such as photo contests and themed virtual meetups, which incentivized users to interact more actively with the platform.

As an intern in the AI department of a major manufacturing company in China, I gained valuable experience in the application of deep learning technologies for industrial automation.

• YOLOv5 Deep Learning Model Training: Trained an AI deep learning model using YOLOv5 with PyTorch, based on an image dataset collected from the factory. Exported the trained model into TensorFlow for inference purposes.
• Model Analysis and Performance Evaluation: Analyzed the performance of the deep learning model, generating Prediction-Recall Curves to evaluate its effectiveness. The model achieved a success rate of over 99% in identifying mechanical parts for production.
• Automated Target Detection: Successfully implemented automatic target detection for the conveyor belt system, improving the factory's operational efficiency by accurately detecting and sorting mechanical parts.
• Collaboration and Issue Resolution: Worked closely with software engineers to develop and test application procedures, ensuring system effectiveness and efficiency. Investigated and resolved issues related to application functionality to maintain smooth operations.
• Version Control and Team Collaboration: Utilized Git for version control, managing code changes effectively and ensuring seamless collaboration within the team.