Mayur Amarnath Palavalli

• Built resume scoring feature end-to-end that evaluates a resume against a job description; currently used by 6000+ subscribers.
• Leveraged vectorized words to build AI model that discerns verbs at the beginning of sentences.
• Used Python, OpenAI, and prompt engineering.

Measuring Variance in LLMs using Symbolic Execution (on-going)

• Goal of this research is to measure the semantic variance between code generated by a given LLM with the same prompt used multiple times.
• Measuring semantic variance via Satisfiability Modulo Theories (SMT) solver.

Improving Syntactic Correctness of LLM-generated Infrastructure as Code(IaC)

• Improved the syntactic correctness of AWS CloudFormation templates generated from LLM by using feedback loop based on error messages produced by cfn-lint.
• Wrote research paper with Dr. Mark Santolucito from Columbia University - accepted to International Journal of Secondary Computing and Applications Research for publication in December 2024.

Facial Recognition App

• Created iOS app with built-in facial recognition algorithm based on regression model created using Convolutional Neural Networks and euclidean distances.
• Built functionality to pair the app with home security systems.

Bluestamp Engineering

• Built self-stabilizing gimbal platform for a camera.
• Built and programmed circuit consisting of Arduino, MPU6050, servo motors, and gyroscopes.
• 3D printed parts and assembled using power tools,

InspiritAI - Virtual Self-Driving Car

• Created a software car that can mimic a human driver to drive itself in an environment that consists of road and grass.
• Leveraged numpy, sklearn, pandas, and tensorflow.
• Implemented reinforcement learning using Q-learning algorithm.

InspiritAI - Facial Emotion Detection

• Created a model that identifies the emotion of a person given a picture.
• Leveraged euclidean distances, KNeighborsClassifier, LogisticRegression, and DecisionTreeClassifier algorithms.
• Tested accuracy of each algorithm using classification trees.