Hello, I'm Ramandeep Singh

I am a Data Scientist with expertise in Deep Learning and Machine Learning.

PROFESSIONAL EXPERIENCE

Here is an overview of my professional experience:

Generative AI Intern Intern – Epilepto - Remote (May 2025 – July 2025)

Developed modular, memory-augmented LLM pipelines using DSPy and mem0, enabling tool-based reasoning, long-term retrieval, and document generation at scale.
Deployed AI solutions integrating model fine-tuning, semantic memory, and structured feedback loops for enterprise use cases.
Achieved a 40% improvement in enterprise workflow accuracy via scalable, feedback-driven agents.
Data Scientist Intern – Parabox -Remote (Oct 2024 - Jan 2025)

Engineered a hybrid ensemble model that integrates a finely tuned XGBoost and CNN, leveraging a weighted probability averaging technique to boost predictive performance and achieve an accuracy of 97%.
Implemented advanced image pre-processing methods, including median blur and Total Variation (TV) Denoising via OpenCV, to enhance image quality and reduce noise in medical imaging data.
Facilitated cross-functional collaboration between technical and medical teams through bi-weekly meetings.

PROJECTS

1) Built a conversational AI chatbot for JUIT, using Hugging Face API and Flask backend. The bot leverages a dataset of 534 training intents and spaCy to craft context-aware responses.
2) Optimized the training pipeline for low-resource systems, achieving an overall accuracy of 94%.

Tech used: spaCy, Flask, Hugging Face API, NLP.

Try it Yourself

Image Dehazing

1) Achieved 72% accuracy in image dehazing using a single FFA-Net model.
2) Designed a 3D attention mechanism to enhance feature extraction, significantly improving visibility restoration.

Tech used: Python, React.js, Flask, OpenCV, PIL, Pytorch, Keras.

Check

Churn Prediction

1) Customer churn is a critical issue for businesses, and analyzing churn patterns helps in developing retention strategies. This project consists of two key parts: 1) Churn Analysis
2) Churn Prediction

Tech used: Tableau, Logistic Regression, Random Forest, KNN.

Churn Analysis
Churn Prediction

GDP Prediction

1) Forecasted national GDP using MySQL (for data warehousing) and XGBoost (for regression).
2) Achieved R² score of 0.9522 on unseen economic data.
3) Streamlit dashboard allows interactive exploration of trends and simulation of GDP outcomes.

Tech used: MySQL, XGBoost, Streamlit, Pandas, NumPy, Matplotlib.

Check

Movie Recommendation System

1) Developed a movie recommendation system utilizing collaborative and content-based filtering methods; integrated real-time data via Web API, resulting in an impressive precision rate of 0.88 while personalizing user interactions with tailored content.
2) Designed a modular data pipeline for preprocessing and vectorizing movie data with CountVectorizer and cosine similarity, resulting in improved recommendation accuracy and efficiency.

Tech used: MYSQL, NLP, webAPI.

Personal Webpage

My own responsive personal webpage showing my skills, projects, education and GitHub and LinkedIn links on Netlify.

Tech used: Netlify, HTML, CSS, JavaScript.

Fake News Classifier

1) Built a fully connected PyTorch-based neural network achieving 91% accuracy, leveraging a custom TF-IDF vectorizer trained on preprocessed text data to extract meaningful features.
2) Designed and deployed an interactive Flask-based web application integrated with MySQL, enabling efficient storage of predictions, user feedback, and queries.

Tech used: Flask, MYSQL, NLP, Pytorch.

IPL Score Prediction

Batting team, bowling team, overs, wickets fell, etc have to be input to get the final predicted score at any time of the game to predict score.

Tech used: Decision Trees, Linear Regression, Random Forest, Lasso Regression, Support Vector Machine, Neural Networks.

Stopwatch

A basic stopwatch made using HTML, CSS, and JavaScript.

TIC TAC TOE GAME

A basic TIC TAC TOE game.

Coin Toss

Coin Toss where the coin revolving animation is there.

Research Experience

Attention Mechanism Based Novel Multi-Branched Approach for BiLSTM Model

Jun 2024 – Present

Developed a novel BiLSTM model with a multi-branch architecture, enhancing accuracy by 3% over recent state-of-the-art methods in multiple applications, in collaboration with professor Dr. Pardeep Garg.
Currently authoring a paper on this approach for submission to IEEE Transactions on Neural Networks and Learning Systems (Impact Factor of 10.2).

ACHIEVEMENTS

Some of my notable achievements:

MakeMyTrip (MMT) myLaunchpadC - Aug 2025
- View Certificate
- Reached final round of MakeMyTrip (MMT) myLaunchpad and was among top 150 out of a pool of 23,000+ applicants and won a cash prize.
Clear Vision Challenge - Oct 2024
- View Certificate
- Led a team in a prestigious one-week AI event organized by the 'Centre of Excellence in Artificial Intelligence for Education.'
- Secured 3^rd place among top-tier participants, showcasing technical expertise, leadership, and creativity.
AI - Boon or Bane - Mar 2024
- View Certificate
- Achieved 3^rd position in an on-site debate, engaging in profound discussions with 20 peers on the ethical and societal implications of AI.
Innovate with NVIDIA - Nov 2023
- View Certificate
- Developed an IoT-Based Water Quality Estimator with an impressive 94.8% accuracy, leveraging Logistic Regression and Oxidation-Reduction Potential (ORP) analysis.
CGPA Felicitation
- View Certificate
- I am honored to have been one of the top 10 students from the CSE program in my year, recognized for maintaining an exemplary 9+ CGPA. This milestone was celebrated during Freshers' Night by our esteemed Vice Chancellor, Prof. Rajendra Kumar Sharma, and Registrar, Rakesh Bassi.
- This acknowledgment not only underscores my dedication to academic excellence but also inspires me to pursue even greater milestones in my technical journey.

EXPERIENCE

Where I have volunteered in my College:

JYC

Coordinator

Sep 2022 - PRESENT

Overseeing educational initiatives for 24 underprivileged students and organizing workshops.

SIAM JUIT

Senior Member

Mar 2023 - PRESENT

Coordinating technical events to enhance engagement within the tech community.

UBC [United Bhangra Crew]

Senior Member

Sep 2022 - PRESENT

Showcasing cultural heritage through performances.

University Grants Commission (UGC)

UGC Saarthi

Jun 2023 - Oct 2024

Guided students in navigating NEP 2020 initiatives.

TECH

Technologies / Tools I like working with:

Deep Learning Python Machine Learning TensorFlow Keras PyTorch OpenCV GitHub SQL JavaScript Docker

EDUCATION

MY Education

B.Tech. in Computer Science & Engineering Jaypee University of Information and Technology, Solan 2022 - 2026 | CGPA: 9.19 (till 6th semester)

Class 12th Guru Teg Bahadur Public School, Faridkot, Punjab 2020 - 2022 | Percentage: 82.80

Class 10th Dasmesh Public School, Faridkot, Punjab 2007 - 2020 | Percentage: 93.60

CERTIFICATIONS

Machine Learning with Python - IBM

Database Structures and Management with MySQL - Meta

Python Foundation - Infosys Springboard

Introduction to Data Science in Python - University of Michigan