About Me

My introduction

Results-driven data scientist with big data, machine learning, deep learning, and GEN-AI expertise. Proficient in Python, Big data technologies, PyTorch, and LLMs with a strong background in developing data-driven solutions across various domains, including healthcare, e-commerce, social media, and IoT. Proven ability to design and implement data analysis, NLP, KG, DL, and CV models. Demonstrated success in hackathons, achieved prestigious scholarships, conducted impactful research, and contributed to open-source projects, showcasing strong problem-solving, collaborative, and analytical skills with a commitment to innovate and impact using data science.

Skills

C/C++ Python Java OOPs Data Structures and Algorithms (DSA) NumPy Pandas Seaborn Matplotlib Scikit-learn PyTorch TensorFlow Machine Learning Deep Learning Data analytics Data visualization OpenCV Langchain Streamlit LLM SQL PostgreSQL Apache Airflow PySpark Docker Shell GCS Minio Big data SPARQL Cypher Neo4J GraphDB OrientDB

Experience

MobilityDB

Open Source Developer

Project Link | July–September 2024 | Brussels, Belgium
  • Improved JMEOS, Java binding for the MEOS spatiotemporal library
  • Technologies: C, Java, FFI, CI/CD, GitHub Actions, Python
  • Contributed 30K+ lines of code to JMEOS and MobilityDB repositories
  • Boosted testing coverage by 70% using JUnit for MEOS data types
  • Automated documentation deployment using GitHub Pages, streamlining API visibility for 500+ users
  • Built CI/CD pipelines with GitHub Actions, cutting build and integration times by 30%

Health Technologies Lab (HTL), IBME, University of New Brunswick (UNB)

Research and Development Intern

Project Link | May–August 2023 | Fredericton, Canada
  • Worked on Translating Foot Pressure Maps to 3D Human Poses
  • Technologies: Pytorch, Python, Mediapipe, TensorFlow, Keras, OpenCV, MATLAB
  • Captured foot pressure maps using 100Hz tiles; mapped to 3D poses with 33 keypoints
  • Used video from 8 cameras as supervision; developed Encoder-Decoder, CRNN, and CNN+LSTM models
  • Evaluated models using MPJPE and MSE, enabling non-invasive person identification with 95% accuracy

Projects

Completed

Clients

Experience

Get in touch

Do you have a project in your mind, contact me here

Find Me

Email: arijit.samal@student-cs.fr

Tel: +32 493139020

Paris, France