• Acted as lead engineer in a team of 4 to build an application using Lastmile AI’s framework, AIConfig. The app was presented to the Lastmile AI team, and intended to be used for marketing and feedback purposes.

• Leveraged AIConfig to use and manage multiple LLM models. Configured model settings, prompts, metadata, and parameters separately from application logic.

• Integrated RAG by segmenting then ingesting a document into a ChromaDB vector database, and creating functions to query from the collection and serialize the data.

• Led 6 team members and derived the technical design for a project targeting an external client application to sync the completion status of user onboarding tasks to an internal checklist.

• Reduced shipment exceptions by 40% by improving the internal user workflow to reduce manual task completion; thus, saving ~15 hours a week in manual work and improving worker productivity.

• Reduced average client time to verify payment from 2 weeks to 4 days and decreased the percentage of blocked customs bond purchases due to unverified payment methods by 25%.

• Ideated a solution to automate invoice CSV creation by consolidating data from multiple sources to enhance processing speeds, and using Looker to provide a dynamic and easy to use interface for non-technical users.

• Integrated in-house AB testing solution that saved the company $36000 per year, and 2 hours per week of engineering time.

• Revamped public website to conversion rates by 5% and average visit duration by 50%.