Mandeep Singh Profile Picture

Hello, I'm

Mandeep Singh

Data Scientist

Get To Know More

About Me

Experience Icon

Data Science Intern

Rally Vision
Sept-Dec 2024

Education Icon

M.Tech (Computation and Data Science)

IISc Bangalore
8.4/10 CGPA

Education Icon

B.E. (Chemical Engineering)

BITS Pilani KK Birla Goa Campus
8.03/10 CGPA

I'm a Data Scientist with a background in Chemical Engineering and an M.Tech in Computational and Data Science from IISc Bangalore.
My expertise lies in building end to end Machine Learning systems, from custom feature engineering and model development to containerizer deployment on cloud infrastructure.
Interested in applying AI to solve real-world problems and creating impactful solutions that are not just accurate, but scalable and production ready.

What I Work With

Technical Skills

Languages

Python SQL

Machine Learning & Data Science

Scikit-learn Pandas NumPy

Deep Learning

PyTorch TensorFlow (Keras) Transformers

Natural Language Processing

Hugging Face TF-IDF Text Embeddings

Audio & Signal Processing

Librosa YAMNet Embeddings

Deployment & MLOps

FastAPI Docker ONNX AWS EC2

Web Scraping & Retrieval

Scrapy Beautiful Soup Chroma DB OpenAI API

Tools

Git

Browse My Recent

Projects

Fungi Image Classification

Built an image classifier using EfficientNet v2, fine tuning the last two blocks and the custom classifier head.
After training, converted the model to ONNX for fast inference on a resource constrained EC2 instance.

Movie Reviews Sentiment Analysis

Trained a DistilBERT-based sentiment classifier using Hugging Face Transformers.
For deployment, used a model trained using TF-IDF + Logisitic Regression due to significantly lower computational requirements.

Music Genre Classification

Built a music genre classification system using YAMNet embeddings and multiple ML models.
For live deployment, used a lightweight Logistic Regression model.

CUDA Runtime API RAG Assistant

Built an intelligent assistant for NVIDIA CUDA Runtime API documentation using Retrieval-Augmented Generation (RAG).
Implemented hybrid retrieval with dense embeddings + BM25, Reciprocal Rank Fusion (RRF), and Cohere reranking for improved answer relevance.

Get in Touch

Contact Me