// AI & GenAI Engineer · Kerala, India

Rijin
Shaji

Building intelligent systems with LLMs, RAG pipelines, and semantic search. Turning unstructured data into production-ready AI solutions.

View Projects Get in Touch Download CV
Scroll to explore
01 /

About Me

Rijin Shaji

I'm an AI & Machine Learning Engineer with a strong foundation in data-driven problem solving and model development. I specialize in Generative AI — building LLM-powered applications, RAG systems, and semantic search pipelines that solve real enterprise challenges.

My background spans from robotics & automation engineering to cutting-edge GenAI, giving me a unique systems-thinking perspective when designing AI architectures. I manage the full lifecycle: data ingestion, feature engineering, model training, deployment, and continuous optimization.

Currently building production-grade AI solutions at Coderzon, leading projects ranging from banking document intelligence to AI-powered recruitment systems.

6+
Projects Shipped
2+
Years Engineering
LLM
RAG · Embeddings · NLP
B.Tech
Robotics & Automation
02 /

Technical Skills

AI / ML / GenAI
LLMs Prompt Engineering RAG Semantic Search NLP Text Classification Embeddings CNNs Transfer Learning
Languages & Libraries
Python SQL NumPy Pandas Scikit-learn TensorFlow Hugging Face FAISS Sentence-Transformers
Tools & Platforms
GROQ API LLaMA 3.3 Git AWS (Basic) MySQL Tableau Jupyter Notebook Google Colab
Core Practices
Feature Engineering EDA Model Evaluation Pipeline Design Hallucination Reduction Vector Retrieval Modular Architecture
03 /

Experience

AUG 2025 — PRESENT
GenAI Engineer – Applied Projects
Coderzon
  • Built LLM-powered applications using LLaMA 3.3 via GROQ API for resume analysis, document understanding, and intelligent query handling.
  • Designed embedding-based semantic search pipelines using Hugging Face models and FAISS for high-accuracy information retrieval.
  • Developed RAG systems for enterprise use cases including banking document analysis and policy-compliant Q&A.
  • Implemented resume–job matching logic combining semantic similarity scoring and rule-based ranking.
  • Structured modular AI pipelines separating ingestion, embedding, retrieval, and LLM prompting for scalability.
AUG 2023 — AUG 2024
Robotics & Automation Engineer
Pacific Weld Systems Pvt. Ltd. · Chennai, India
  • Installed and commissioned industrial robotic and PLC-based systems across manufacturing environments.
  • Executed end-to-end system integration including calibration, signal mapping, and controller configuration.
  • Diagnosed and resolved system-level failures — skills directly transferable to large-scale AI system debugging.
04 /

Projects

PROJECT 01
AI-Powered Job Matching Platform
Full-stack LLM-powered recruitment platform using semantic resume–job matching. Features FastAPI backend, MySQL database integration, automated resume parsing, experience calculation, and relevance ranking — replacing keyword-based matching with deep semantic understanding.
LLaMA 3.3 GROQ API Hugging Face FAISS FastAPI MySQL NLP Python
GitHub
PROJECT 02
Enterprise RAG System — Banking
Team Lead. Designed an enterprise-grade RAG pipeline for secure banking document analysis covering KYC validation, fund transfer policies, and regulatory compliance queries.
RAG LLaMA 3.3 70B Sentence-Transformers FAISS GROQ API
GitHub Live Demo
PROJECT 03
Facial Expression Recognition System
Real-time emotion classification system using deep learning. Built a live prediction pipeline for camera feeds with MobileNetV2 transfer learning optimized for limited hardware.
MobileNetV2 CNN Transfer Learning TensorFlow OpenCV
GitHub
PROJECT 04
Handwritten Digit Recognition
Interactive GUI application where users draw digits on a canvas and receive real-time CNN predictions with confidence scores. Features automatic digit cropping, centering, and image preprocessing — trained on MNIST dataset.
CNN TensorFlow Keras Tkinter NumPy Pillow MNIST
GitHub
PROJECT 05
Iris Flower Species Classifier
End-to-end CNN pipeline classifying Iris flower images into Setosa, Versicolor, and Virginica. Includes dataset splitting, image augmentation, multi-class CNN training (32→64→128 filters), and single-image prediction with confidence scores.
CNN TensorFlow Keras OpenCV ImageDataGenerator NumPy Python
GitHub
PROJECT 06
Leaf Disease Detection System
Computer vision system for early crop disease detection. A custom CNN with binary classification distinguishes healthy from infected plant leaves — featuring dataset splitting, image augmentation, model training, and single-image inference for real-world usage.
CNN TensorFlow Keras OpenCV NumPy Computer Vision Python
GitHub
05 /

GitHub Activity

Mon
Wed
Fri
— contributions in the last year
View GitHub Profile
Less
More
06 /

Get In Touch

Ready to build something amazing together?

Have a GenAI project, collaboration idea, or just want to say hi? Drop me a line.

LinkedIn GitHub