Pranay Narula

I Build ML Algorithms

Undergraduate Student | Machine Learning Enthusiast

Started Coding

Age 13 (2018)

Discovered the world of programming and wrote my first lines of code

ML Journey Begins

Age 15 (2020)

Started learning machine learning and built my first neural network

Startup Journey

Age 17 (2023)

6 months of algorithm development, 2 months of prototype building

  • Algorithm Development (6 months)
  • Prototype Building (2 months)

Work Experience

Team Lead Intern

Redvest

January 2025 - Present

  • Supporting data-driven decision-making through statistical analysis and data visualization.
  • Collaborating with cross-functional teams to identify trends and actionable insights in the UI.
  • Working with large datasets to provide actionable solutions for growing revenue
numpy
firebase
matplotlib

AI Safety Fellow

Uconn Beacon AI

September 2024 - November 2024

  • Engaged in in-depth paper reviews, including implementations and activities centered around AI safety and ethical considerations.
  • Collaborated with peers and mentors on understanding the overarching risks of AI and mitigating them
  • Developed an understanding of safety protocols and frameworks applicable to machine learning and AI models.
JS
Python
matplotlib
firebase
numpy

InterviewAI Intern

InterviewAI

May 2024- August 2024

  • Developed v0 of InterviewAI Async, a B2B SaaS for InterviewAI using NextJS, OpenAI, and Supabase
  • Developed the entire payment integration and subscription system using hooks and Stripe
  • Gained hands-on experience in transforming user specifications into code, and writing good copy
OpenAI API
Supabase
Nextjs

Education

Bachelor of Science in Computer Science

Concentration: Algorithms

University of Connecticut, Storrs, Connecticut

2024 - 2028

Relevant Coursework

Algorithms and Complexity
Multivariable Calculus
Data Structures
Discrete Math
Statistics
Intro to ML Kaggle

Featured Projects

FelicityAI - Startup for Efficient LLM Fine-Tuning using Explainability

FelicityAI helps fine-tune large language models (LLMs) by leveraging an explainability algorithm designed implemented and tested by me. It determines the contribution percentages of different categories in datasets, enabling better insights and improved model performance.

HuggingfaceNext.jsFastAPIPyTorchPEFTDatasetsTransformers

ASL Teacher Using Image Recognition and Spaced Repition built for HackMHS

Learn ASL using spaced repition through a convolutional Neural Network

Convolutional Neural NetworksAlgorithmsAPI Integrations

Reinforcement Learning Book

MiniBook On RL

pythonRLLatex

Adversarial Attacks on Face Recognition (Hackathon Winner)

Implemented a Projected Gradient Descent (PGD) adversarial attack on facial recognition systems using the FaceNet model to demonstrate vulnerabilities in high-security environments like airports.

pythonPGDFaceNet

Skills

Machine Learning

  • Neural Networks
  • Deep Learning
  • Computer Vision
  • NLP

Programming

  • Python
  • PyTorch
  • TensorFlow
  • scikit-learn

Tools

  • Git
  • Docker
  • Linux
  • Jupyter

Math & Stats

  • Linear Algebra
  • Calculus
  • Probability
  • Statistics