▪ Led and built Otter’s first Meeting Types, Rules Engine, Workflow, and AI Classification systems from 0 to 1, enabling new automated insight and action pipelines. Blog post: https://lnkd.in/dEEfPApH

▪ Enhanced Otter's insights/action items/summary/outline systems.

▪ Scaled the weekly digest/summary email systems.

▪ Developed a web application for an HR Platform with React.js and deployed with AWS (EC2, S3).

▪ Built an authentication backend (10+ endpoints total) for 2000+ users to control user access using Node.js and FastAPI with JWT and OAuth2, storing data in PostgreSQL and deployed through AWS RDS, API Gateway, Lambda, and EC2.

▪ Designed the data model and constructed the backend (20+ endpoints total), piped from Indeed, for company's recruiting portal with FastAPI and SQL, deployed with AWS RDS.

▪ Processed data (5000+ entry per year) from the financial database and performed financial analysis on categories including industries, products, R&D, and functions based on its half-year and full-year data for two weeks duration per six months.

▪ Reduced 640+ hours in total for management team by conducting a detailed analysis of the customers for more than 200 companies, concluding the trend of the performance of the sales, and aggregating data to have a greater scope of the sales performance for the whole year and the importance of the customers.

▪ Developed a smart database backed with SQL and enabling data flow with Airflow, built Flask function to optimize the querying and analytic process.

▪ Designed the PowerPoint presentation and present the data analysis to help the sales team visualizing their performance in the last season and tasks for the next sales season.

▪ Designed and Constructed company’s chatbots APIs, developing with FastAPI, fine-tuning with PyTorch on RunPod, storing data in SQL and Pinecone (vectorDB) for backend, and deploying with Docker on Azure (Hosting database server), integrated CI/CD.

▪ Constructed company’s data pipeline API that converts customers' data into structured data storing in SQL database and vector database, leveraging the power of Large Language Models, developed in FastAPI.

▪ Reviewed research papers on large language models’ performance enhancement, applying state-of-the-art techniques, such as in-context learning, fine-tuning, Chain-of-Thoughts, and prompt-engineering, to enhance chatbot interactions on self-constructed testing platform.