Kai Avni

Robotics & Artificial Intelligence Laboratory

• Developed the first AI assistant for a new product using retrieval-augmented generation (RAG), with the ability to answer questions and assist with use cases (e.g. filling out API payloads). • Delivered a fully functional proof-of-concept, along with comprehensive project documentation and next steps. • Technologies used: Python, FastAPI, Docker, PostgreSQL, Azure, OpenAI embedding and GPT-4 models, Redis.

• Updated critical .NET industrial applications in C# to write to a new SQL database. • Published a training guide of 50+ pages for common issues on manufacturing lines that was used to train production associates (made of 50% new hires).

Robotics & Artificial Intelligence Laboratory • Implemented (MATLAB/C++) an autonomous navigation program capable of path following and obstacle avoidance using parametric trajectory generation, gradient descent, and forward models. • Tested and validated this control algorithm in simulation (ROS/RViz) and on a physical mobile robot.

• Launched and developed a global product offering for Megapacks by gathering customer feedback and aligning cross-functional teams. • Saved $14.5M by identifying a solution to meet the needs of Tesla field service technicians and minimizing cost and complexity to the customer. • Authored and published (15+ pages of) customer-facing product documentation. • Designed infrastructure CAD models, using a rapid prototyping and user-centric approach.

• Developed an internal workflow process to systematically identify "Bundling Opportunities," resulting in $200k/year savings for the company and minimizing customer turbine downtime. • Developed a standardized report format for the company's inspection suppliers to follow when delivering reports, speeding up the review time for engineers within the company. • Managed retrofit campaigns for field quality issues affecting major wind turbine components; coordinated between site managers, engineers, technicians, and suppliers.

The Solar Decathlon Design Challenge is a yearly competition to design a sustainable home. Team accomplishments: • Designed a 4-bedroom house with a capacity of up to 5 occupants. • Achieved an HERS rating of 10 and a net construction cost of $109,000. • Qualified as a finalist team (top 10). • Partnered with Habitat for Humanity and a local Community Solar Project. • Developed a 48-page Final Report and a 2-page Project Summary (attached). The UR Solar Decathlon Team consisted of 10 students across 3 different engineering and analysis sub-teams (Energy, Mechanical, and Market Analysis). Full project portfolio: https://drive.google.com/drive/folders/1QT9S3JKYOjQcRAotzCHXexMl6tOsKLIu?usp=sharing

I launched my own tutoring business called Kai Academy (a quip on Khan Academy) for students in the community. Over the course of 2 years, I tutored several students from different local schools in math and science. I helped with homework, exam-prep, study-skills, and motivation (growth mindset). Students experienced increases in their scores and confidence; I was also recommended by the teachers and the school principal.

E5 Scholar | Resident Advisor | ASME Scholarship | SHPE Scholarship | HSF Scholar | Intel Scholarship | Berkshire Hathaway Energy Scholarship