About Me

I am a Machine Learning Engineer currently pusuing my Bachelor's in Technology in Electronics and Instrumentation Engineering at Odisha University of Technology and Research, Bhubaneswar. Considering my professional experience,I have completed my internship as a Data Science Intern at Analytic Labs, an AI developer at the Agritech startup- Farmoid Robotech Pvt. Ltd. and a Student Research Intern at the Centre of Excellence in AI(Tech Mahindra,OUTR). I have worked on my personal projects like "Fine-Tuning of Yolo-v8 for vehicles detection in a traffic footage", "Text-Recognition and Keyword Searching using GOT-OCR model" and "Diabetis Detection using AI" which I have containerised using Docker and deployed it on FastAPI. I love solving real-world problems with the help of trending technologies like Machine Learning/Artificial Intellgence, MLOps, IoT etc.I have an indermediate foundation of Data Structures and Algorithms and I solve them using Python.

Resume

Experience

Data Science Intern @Analytic Labs

(Jun,2023 - Sept,2023)

Successfully directed and delivered three research-based projects—Cloud Instance Forecasting, Demand Forecasting Using Time Series, and Diabetic Detection Using AI, providing innovative and data-driven solutions to clients.
Analyzed historical usage patterns of Cloud Computing resources and applied advanced machine learning algorithms to predict future resource demands.
Utilized Time-Series analysis techniques, such as ARIMA and LSTM models, to predict future demand trends for various products and services that allowed clients to anticipate market changes and plan inventory effectively.
Assisted clients in writing and formatting research papers, ensuring high-quality documentation.
Experimented with various FaaS methods, improving the Forecasting Accuracy by 30%.

Artificial Intelligence Developer @Farmoid Robotech Pvt. Ltd.

(Feb,2024 - Jun,2024)

Spearheaded a 3-member team in developing AI/ML-methods for real-time operations, resulting in a 30% accuracy improvement in areas like Crop Yield Prediction, Market Price Prediction, etc.
Designed and implemented an innovative Crop Recommendation system leveraging the weather data collected in real-time to enhance the decision-making process in farmers.
Successfully deployed the above-mentioned projects on Streamlit, ensuring seamless integration and user-friendly access.

Student Research Intern @Centre of Excellence in AI (OUTR), Tech Mahindra

(Jan,2024 - Present)

Lead a research project focused on tuning PID Electronic Controller parameters using various Optimization techniques, aiming to improve their Accuracy by 20%.
Actively reading, reviewing, and implementing findings from latest research papers in Control Systems and Machine Learning, applying the latest methodologies.

Projects

Diabetis Detection using AI

Developed an AI-based application to detect Diabetes by analyzing medical data and predicting conditions with high accuracy.
Made a comparative analysis of ML models like Random Forest Classifier, Decision Tree Classifier, Naive Bayes Classifier, Support Vector Machine Classifier and Logistic Regression to find the best model for classification.
Evaluated the model's performance on the basis of metrics like Accuracy, Precision, Recall, F1 Score, Confusion Matrix and ROC-AUC Curve.
Containerized the application using Docker and deployed it on FastAPI for robust and scalable accessibility.
The application can also classify the Diabetis as Type-1/Type-2 other than detectiing its presence/absence.

View on GitHub

Fine-Tuning Yolo-v8 for vehicle detection,tracking and classification

Fine-Tuning of Yolo-v8 for vehicle detection,tracking and classification

Developed a real-time traffic monitoring system by fine-tuning YOLO-v8 for accurate vehicle detection, classification, and tracking in a traffic video footage.
Used a custom traffic video footage as the dataset for tuning the model's state-of-the-art object detection capabilities for a particular task of vehicle detection and classification
Utilized the Supervision library of Roboflow to process and analyze traffic data to ensure the correct prediction of the vehicles as per their Confidence Scores.
The combination of the Supervision library and the Yolo-v8 helped in the improvement of the prediction Accuracy upto 90%.

View on GitHub

Drug Reviews Classification using NLP

Built an AI-driven system to classify drug reviews into categories like positive, negative, or neutral, aiding in analyzing patient sentiments and drug efficacy.
Employed advanced Natural Language Processing (NLP) techniques to preprocess, analyze, and classify textual data from user reviews.
Conducted a comparative analysis of various text classification algorithms, including Passive Aggressive Classifier and Naive Bayes Classifier etc.,to achieve optimal Accuracy.
Designed and deployed the project using Gradio interface, providing an user-friendly interface for real-time drug reviews classification.
Containerized the application using Docker and deployed it on FastAPI for robust and scalable accessibility.