Arijit R.

Key Skills

Software

OpenCV AI Kit (OAK)

Don't disclose

Top Subject Matter

Optical Character Recognition (OCR) for Chakma script

character recognition model training.

Real-time communication system engineering (message/data handling).

Top Data Types

Document

Image

Top Task Types

Classification

Data Collection

Transcription

Freelancer Overview

Summer Intern, National Informatics Centre (NIC) — OCR system engineering for a low-resource language.. Brings 2+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include OpenCV AI Kit (OAK) and Don't disclose. Education includes Bachelor of Technology, National Institute of Technology, Agartala (2026) and N/A, Umakanta School (2022). AI-training focus includes data types such as Document, Computer Code, and Programming and labeling workflows including Classification, Data Collection, and OCR.

Labeling Experience

Summer Intern, Computer Vision (OCR) - National Informatics Centre

ImageClassificationTranscription

Worked as a Summer Intern to build an OCR system for the Chakma script, focusing on digital preservation and accessibility. Implemented and trained a convolutional neural network using Python with TensorFlow and OpenCV to classify characters. Created a data augmentation pipeline to expand a limited dataset and improve model robustness and generalization. • Designed and implemented an OCR pipeline tailored for a low-resource script • Trained and evaluated CNN models using TensorFlow and OpenCV • Built data augmentation (rotation, scaling, noise injection) to increase dataset size • Assessed performance on a validation set to achieve high character recognition accuracy

2025 - 2025

Summer Intern, National Informatics Centre (NIC) — OCR system engineering for a low-resource language.

OpenCV AI Kit (OAK)DocumentClassification

Built an OCR pipeline for the Chakma script to recognize characters and support digital preservation and accessibility. Trained and evaluated a CNN-based model to classify characters using TensorFlow and OpenCV. Implemented dataset expansion via augmentation (rotation, scaling, noise injection) to improve recognition robustness and generalization. • Performed character recognition training and validation to exceed 92% accuracy. • Designed an OCR workflow tailored to a low-resource writing system. • Created a data augmentation pipeline to grow the dataset by 300%. • Focused on improving model robustness through controlled perturbations.

2025 - 2025

Winter Intern, Oil and Natural Gas Corporation (ONGC) — MERN real-time chat application development.

Don't discloseData Collection

Developed an internal real-time chat system supporting concurrent users, enabling continuous message exchange. Implemented persistent low-latency communication using WebSockets via Socket.IO to ensure timely delivery. While not traditional data annotation, the work involved engineering data/message handling flows for real-time system functionality. • Built the application using the MERN stack (MongoDB, Express.js, React.js, Node.js). • Implemented WebSocket-based communication for up to 50 concurrent users. • Focused on low-latency, persistent message delivery behavior. • Contributed to backend and frontend integration for real-time chat.

2024 - 2025

Education

Work History