Ayush Patne

Machine Learning · AI · Products · Exploration

Email LinkedIn GitHub

About

I work across Machine Learning, AI, and product experimentation. Currently pursuing my MSc at Leeds, I'm fascinated by how LLMs and generative models work, and would love to be part of building and improving them.

I've built Hopps, a social dining project aimed at making group plans effortless with ML-driven recommendations, and Maira, an experimental platform exploring emotional intelligence in AI and AI Safety. I also completed a research internship at IIM Udaipur, where I developed data pipelines and applied ML for marketing and NFT research.

I'm looking for opportunities in ML engineering, LLM training, or applied AI research, anything that blends technical problem-solving with creative exploration.

Education

MSc Advanced Computer Science (Artificial Intelligence)

University of Leeds · Sep 2025 – 2026

B.Tech. Artificial Intelligence & Data Science

Vishwakarma Institute of Information Technology, Pune · Dec 2021 – May 2025

Experience

Founder · Hopps

Sep 2024 – Jun 2025 · Pune, India

  • Led cross-functional team of 8+ building social dining platform connecting users with 70+ restaurants
  • Designed ML-based recommendation engine using Flutter, Firebase/GCP, and Gemini API
  • Analyzed restaurant and user data to inform product strategy and feature prioritization

Research Intern · IIM Udaipur

Sep 2023 – Apr 2024 · Remote

  • Built automated classification system processing 250,000+ NFT records using Hugging Face
  • Developed web scraping infrastructure gathering 113,000+ pages for market analysis
  • Contributed to two research manuscripts using Doc2Vec similarity algorithms

AI Intern · Konverge.AI

Aug 2023 – Oct 2023 · Nagpur, India

  • Applied ML techniques for customer segmentation and purchase prediction
  • Built sentiment classification model using NLP to analyze user review data
  • Developed preprocessing pipelines and visualization for model performance

Projects

BrewStories - LLM Fine-tuning for Creative Text Generation

  • Fine-tuned TinyLlama (1.1B params) on 10K-story dataset using PyTorch & Hugging Face
  • Built end-to-end training pipeline achieving stylistic consistency in creative content
  • Runner-up at Visionary TechFest 24 AI/ML hackathon (in association with Binghamton University)

Distributed Training Infrastructure for Fraud Detection

  • Architected federated learning system across 3 distributed clients for privacy-preserving training
  • Achieved 99.7% AUC-ROC and 96.6% recall on Ethereum transaction analysis
  • Built Flask visualization dashboard for monitoring model performance metrics
  • Bachelor's dissertation project with detailed technical documentation
Read The Blog →

Maira - AI Companion System

  • Engineered conversational AI integrating Gemini 2.0 API with embedding-based memory
  • Built full-stack app using Flutter, Firebase, GetX for emotionally intelligent conversations
  • Implemented context persistence system for long-term memory and emotional support
Try Maira (beta) For Free →

Social Media Recommendation Engine

  • Engineered a TikTok-style recommendation engine using Python, Flask, FAISS, and Google Gemini Embeddings for semantic video retrieval
  • Implemented a vector-based similarity model to understand and recommend content by meaning rather than keywords
  • Optimized personalization through dynamic user embeddings, sliding preference windows, and an exploration–exploitation strategy
  • Deployed an interactive demo platform to simulate real-time user engagement and showcase algorithmic decision-making
Read The Blog →

Technical Skills

AI & Machine Learning Core

Large Language Models Generative AI NLP Transformers Model Fine-tuning Federated Learning RAG Systems

Frameworks & Tools

PyTorch TensorFlow Hugging Face LangChain Python scikit-learn

Product & Infrastructure

Flask Firebase GCP Flutter SQL ETL Pipelines API Development

Research & Publications

"Attention Span and its Correlation with Mental Health"

Richa Shah, Sarthak Rak, Ayush Patne, Dr. Laxmi Bewoor

4th Asian Conference on Innovation in Technology · Aug 2024

  • Led research design analyzing correlation between attention span and mental health across 84 participants
  • Defined analytical approach using KNN clustering for behavioral segmentation