Hello world, My name is
Kayla Ippongi
MS Artificial Intelligence @ Johns Hopkins | Senior SWE @ Rivian
About Me
Hi there! My name is Kayla!
I’m a software engineer with 5 years of experience and an M.S. in AI, and I’m passionate about using machine learning and intelligent systems to solve real-world problems. At Rivian, I’ve built platforms that saved millions and sped up workflows 10×, and in my AI research I’ve worked on computer vision and autonomous vehicle datasets. I’m excited about pushing the boundaries of what AI can do and building products that have a meaningful impact at scale.
My background spans full-stack development, AI/ML systems, and enterprise integrations. I thrive on solving complex technical challenges that deliver measurable business value - whether that's building AI models, architecting scalable GraphQL APIs, or integrating and building microservices.
Recent focus areas:
AI / ML
Backend & APIs
Cloud & Data
Tools & Frameworks
Experience
Senior Software Engineer
Rivian | Sept 2022 - Present
- Led platform integration that automated Fleet workflows—reducing 131 labor hours/day and saving $2M annually
- Optimized API latency from 15s to 1.5s (10x improvement), saving $6M annually in wait time
- Migrated e-signature services to Box API, ensuring feature parity while improving scalability
Software Engineer
Rivian | April 2021 - Sept 2022
- Built backend services for Recall and Quality Containment dashboard, reducing manual tracking across thousands of vehicles
- Improved recall data retrieval speed by 6x using ElasticSearch optimization
Software Engineer
Wayfair | June 2019 - Feb 2020
- Developed Storefront Product Options features, improving engagement across millions of products
- Led cross-functional API migration to serve all product option data
DevOps Developer Intern
IBM | May - July 2018
- Built internal tool to optimize Kubernetes port lookup process for cloud security team
Projects
Computer Vision
YOLOv4 Street Parking Detection
- Curated and annotated a combined dataset from Argo, Waymo, and nuScenes (4,500 frames) and evaluated YOLOv4 vs. Faster R-CNN for on-street parking localization.
- Achieved up to 83% precision (combined mAP ≈ 64%), identified dataset-combination benefits, and proposed extensions (3D detection, sign-reading, knowledge graphs)
Car Lane Detection
- Real-time computer vision pipeline for autonomous vehicle lane detection using OpenCV and Python
- Implements multi-stage image processing including Canny edge detection, Hough line transformation, and region-of-interest masking to accurately identify lane boundaries in video streams under diverse road and lighting conditions.
Natural Language Processing
Machine Learning & AI
Autonomous Racing Reinforcement Learning System
- Autonomous racing system using reinforcement learning algorithms to navigate a racetrack with varying track layouts and obstacles.
- Implemented various reinforcement learning algorithms including Q-learning, SARSA, and DQN to navigate the racetrack and avoid obstacles.
- Features collision detection using Bresenham's line algorithm, intelligent crash recovery with breadth-first search, and performance analysis across multiple track configurations
Writing Quality Prediction System
- Machine learning system for automated writing assessment using keystroke behavioral analysis.
- Developed for Kaggle's writing quality competition, implementing sophisticated ensemble learning with CatBoost, Random Forest, and Linear Regression.
- Features extensive feature engineering on writing process data to predict essay quality from typing behavioral patterns
Research and Writing
Contingencies and Implications of Artificial Consciousness
- Consciousness and alignment in AI models in something that I'm really interested in!
- I explored whether we can create truely conscious AI or just systems that fake it, through analyzing past consciousness frameworks and artifical qualia implementations.