Developed and launched the Otter.ai Desktop App with Electron.js, featuring meeting detection notifications, auto-updates, and secure browser OAuth to deliver a seamless cross-platform experience.

• Enhanced Shift Management apps’ performance by 40% using Node.js, React.js, and state optimization,improving responsiveness and user satisfaction.

• Designed and implemented six core frontend features with React.js, GraphQL, and JavaScript, enabling

real-time taxi roll-call management and improving operational efficiency.

• Led NFC and audio module development with H5 JavaScript, serving 80,000+ drivers in Japan, and integrated Docker and Kubernetes for microservices deployment.

• Built AI-powered automation bots using LLM frameworks LangChain, to generate daily sales reports, reducing manual effort by 90%.

• Integrated Jenkins for CI/CD pipelines, improving deployment efficiency and system security in a microservices architecture.

• Developed Next.js connectors for Shopify and TikTok Shop, handling over 1,000 daily transactions and

streamlining frontend operations.

• Enhanced automation efficiency by integrating 48 apps into Lark via Workato platform, Kafka and Cassandra, contributing to 200+ workflows.

• Collaborated in Agile teams to plan and execute complex microservices-based systems with Redis, PostgreSQL, and Elastic Search, ensuring scalability and high availability.

Identify the spatial cell cell communication from spatial transcriptome data using deep graph neural network

• Developed a deep graph neural net to identify the major cell-cell communication patterns in spatial transcriptome data.

• Discovered key ligand-receptor interactions driving biological processes like cancer progression, providing insights into cell communication mechanisms relevant to drug discovery and disease modeling.

Learn pseudo-spatial-temporal Map in spatial transcriptome via Deep Graph Network

• Proposed a unsupervised deep graph model to learn low-dimensional spatial-temporal relationships of cells in spatial transcriptome data.

• The method reach the state-of-the-art performance in unsupervised segmentation and lineage inference tasks, and has unraveled new biological insights about cancer progression, development, and aging.

Bioinformatic Analysis for DNA methylation patterns and gene expression in mouse hippocampus

• Conducted a bioinformatic analysis of region-specific DNA methylation patterns in the mouse hippocampus, examining associations with global genomic features.

• Characterized sex-specific differentially methylated regions (DMRs) and identified gene targets impacted by age- and sex-related methylation changes, providing insights into epigenetic regulation and gene expression in neural tissues.

Image Super Resolution Research for Super-Resolution Microscopy Images

• Developed Super Resolution models to enhance resolution of microscopy images, improving visualization of fine cellular structures for biological research.

• Applied advanced super-resolution techniques to achieve higher image clarity, enabling more accurate analysis of sub-cellular components in super-resolution microscopy.

• Larksuite Official Website: Launched the first version of the lark official website for global market, implemented 10+ front-end and animation features and A/B testing ability for optimizing user engagement using React.

• Built a user growth data analysis dashboard for Lark’s official website, enhancing decision-making through actionable insights.

Launched a Flashcard App using Swift, featuring vocabulary reminders based on Ebbinghaus’s forgetting

curve, diction and theme management, and a learning stats panel. Achieved 10,000+ downloads and 1,000+ in-app purchases on App Store. (App link below)

Developed a Wallpaper App with 10,000+ downloads and 2,000+ subscriptions, integrating user OAuth, customized wallpaper editing (filters, stickers, etc.), infinite scroll for the wallpaper wall, and seamless image uploads and sharing. (App link below)