// AI/ML Engineer · Gen AI Engineer

Hi, I'm
Cashlin

AI/ML Engineer and Gen AI Engineer turning LLMs and machine learning models into working applications, from raw data to deployed product.

1st Place — ML Empowerment Build Challenge, Productivity Booster Track
View my work Download resume GitHub profile ↗

// 01 — About

Who I am

I'm Cashlin, an AI/ML Engineer and Gen AI Engineer who takes projects from raw data to a deployed product. I build LLM-based applications, machine learning models, and the pipelines that connect them.

Recent work includes a RAG document assistant that answers questions only from uploaded files, and a review classification system built with PySpark MLlib on 6.99M Yelp reviews. I focus on clean data pipelines, honest evaluation with real metrics, and apps people can actually use.

Open to full-time AI/ML Engineer, Gen AI Engineer, and Data Scientist roles, in India or remote. B.E. Computer Science (AI Specialization), 2026.

Programming Languages

Python C SQL R

GenAI & LLMs

LLM APIs RAG Prompt Engineering LangChain ChromaDB Vector Databases

AI & Data

PyTorch TensorFlow Scikit-learn PySpark NLTK OpenCV Pandas NumPy

Tools

Git GitHub Docker Streamlit Hugging Face MS Excel

// 02 — Achievements

Recognition

Hackathon Win

1st Place — Productivity Booster Track, ML Empowerment Build Challenge

Won for LocalPulse AI, a review classification system trained on 6.99M Yelp reviews that reached 86.2% accuracy on a 1.4M-review held-out test set.

// 03 — Experience

Where I've worked

Technical Research Officer

Quantelle Private Limited · Chennai, India · July 2024 – January 2025

// 04 — Projects

Things I've built

02

DocuMind — RAG-Powered Knowledge Assistant

Answers questions from uploaded PDF and TXT files using only retrieved content. Built the pipeline with PyPDF2 extraction, token-based recursive chunking (500 tokens, 50 overlap), and ChromaDB cosine-similarity retrieval of the top 4 chunks. Reduced hallucination with a context-restricted prompt, an explicit fallback when the answer isn't in the document, and low-temperature (0.2) generation.

Python LangChain ChromaDB RAG Streamlit
04

IntelliRead — Book Discovery Engine

Collaborative-filtering recommender trained on 10,000+ Goodreads records, returning personalized top-5 book recommendations.

Python Scikit-learn Pandas Streamlit
04

GoalForge — Intelligent Goal Recommendation System

An AI system that takes your ambitions as input and generates a structured roadmap, suggesting relevant goals, milestones, and resources to make personal and professional growth more systematic.

Python ML NLP

Let's work
together

I'm open to full-time AI/ML and Gen AI roles in India or remote. Happy to talk about any of these projects or the problem you're trying to solve.