I am building VIP Twin, a privacy-first personal intelligence layer designed to store, retrieve, and reason over user context. Designed and implemented a layered memory architecture enabling structured long-term context and personalized interaction. Developed backend systems using FastAPI, MySQL, and Qdrant for hybrid structured and vector-based retrieval. Built custom retrieval pipelines and prompt orchestration to improve context relevance and response quality. Focused on local-first, privacy-preserving AI, exploring trade-offs between on-device inference and GPU-backed deployment. Currently part of the Innovate Bay Scale Sprint incubator, actively building toward early users and product-market fit.

Founding Engineer at Sparkkland, an AI-driven education platform focused on personalized learning and content generation. Building backend systems and AI-powered features to support dynamic content creation, user personalization, and scalable platform workflows. Developing APIs and services using FastAPI and MySQL, with integration of LLMs for generating and structuring educational content across multiple categories. Contributing to product architecture, including token usage systems, content categorization, and teacher/student interaction flows. Collaborating closely with the founders on product direction, feature design, and rapid iteration toward a functional MVP.

Developed and maintained a website using Google Suite, including copywriting, logo design, and layout refinement. Worked directly with the customer to achieve the desired aesthetic and conducted on-site photography of students, staff, and facilities to populate the site. Additionally, designed and implemented a database for managing teachers and students, which included creating a timesheet system and generating QR code badges linked to the backend database for streamlined identification and record-keeping.

Created engaging lesson plans for 7th and 8th-grade students with minor learning disabilities, fostering academic advancement. Adapted teaching methods to diverse learning styles, enhancing student comprehension and participation.

MagiKids is an after-school education camp where I teach classes of six children aged 8-11. I focus on imparting the fundamentals of robotics design, assembly, and programming. Using VEX IQ, VEX GO, and OLLO Robotics Kits, I provide an enriching and hands-on learning experience.

During the VEX IQ season, I successfully coached two teams to become tournament champions. Additionally, I led one team to victory at the Google Girl Power 2023 competition, showcasing our commitment to excellence and innovation in robotics.