Hi, I'm Rudramani,

🧠 → 🤖

Hi, I’m Rudra! I specialize in Machine Learning and Neural Engineering with a background in statistics, computer science, and mathematics. I love taking initiatives and collaborating, whether through applying ideas from different fields or connecting people from different backgrounds.

Even at the moment, I’m merging my two fields - Biomedical Engineering & Information Technology - at Columbia to push the boundaries of Brain-Machine Interfaces.

Education & Experience

For more information, have a look at my curriculum vitae .

  • Columbia University Oct 2022 - Present
    Research Assistant
    Developed new features for AI Model Share MLOps platform, including model deployment, continuous improvement, and ML analytics, as well as automated machine learning model replication toolkit for enhanced reproducibility research.

    Integrated state-of-the-art deep learning models from ML Reproducibility Challenge 2022 using AWS Lambda, resulting in increased efficient deployments and access to ML models.
  • VAO Labs Dec 2021 - Mar 2022
    Machine Learning Consultant
    Led a team of founding engineers to design NLP solutions for supply chain predictability and visibility, yielding high-impact pitches to investors and clients.

    Engineered a decision graph framework for hierarchical classification of over 500 classes, with parallel traversal and Q&A justification.

    Devised techniques for model inspection, logging, and data caching, resulting in an 80% improvement in R&D performance and enabling rapid development of a working prototype within 1 month.
    Dense Passage Retrievers Information Retrieval Systems (IRQA) Text2Text Generation Data Ingestion Abstractive Summarization Framework engineering Model Analysis & Inspections Elasticsearch FAISS T0 T5 BERTs BARTs RoBERTas YOLOv5 custom DNN classifiers
  • NEUROVISION Jun 2021 - Oct 2022
    Founder
    Incorporated a small specialized Machine Learning Research & Consulting Firm with a team of 10 engineers.

    Drove over 10 B2B customer success; Customers include NullIot, Cart Geek, and Axis Technical Group.

    Developed solutions ranging from site optimizations with Google Analytics post-processing to automation using seq2seq chatbots.
  • AZYO Mar 2021 - Dec 2021
    Machine Learning Engineer - Intern
    Designed an authentication system using facial recognition and anti-spoofing countermeasures against presentation attacks.

    Constructed an exam proctoring system detecting head, gestures, and eye movements to generate a confidence metric with RNNs.

    Led development of sign language recognition project from data collection, processing, implementation, to deployment; Achieved 95% accuracy on in-house benchmarks with every batch vocabulary R&D cycle.
    Mediapipe Face-recognition JavaScript Framework Engineering Vector Space Mathematics RNNs LSTMs Video Augmentation Scripting Data Engineering Unreal Engine Django Flask SQL
  • Machine Learning Engineer - Intern
    Refactored entire OCR API pipeline and integrated it with Google Vision API for up to 30% faster performance and better accuracy.

    Created post-processing engine to support 11+ different document formats with Named Entity Recognition and REGEX.

    Managed a team of 6 game developers to design and develop a virtual reality space for multiplayer online education.
    Flask Tesseract OCR EAST Detector Google Vision API spaCy REGEX
  • Master of Science - MS, Bioengineering and Biomedical Engineering
    Won Shardashish Interschool Fellowship, worth $50,000, nominated by Dean of Columbia Engineering.

    Won Data Science Institute Scholar award for Developing MLOps Tools for Researchers.

    Courses: High Performance Machine Learning, Deep Learning for Computer Vision, Competitive Programming.
  • Bachelor of Engineering

    Information Technology
    Undergraduate research assistant to Prof. Sunil Wankhade & Prof. Nilesh Rathode.

    Head Of Department, ABIT-RGIT & IEEE-RGIT; Technical Consultant & Trainer, RGIT-CodeCell.

Supplementary Experience

Further details are available on LinkedIn .

Responsible for managing and co-ordinating 9 technical team members to provide expertise and develop solutions to any technical challenges faced by the ABIT committee.

I provide trainings on how to practically apply deeplearning to solve real-world problems on behalf of codecell RGIT.

I am responsible for satisfying all the technical needs of IEEE-RGIT. I also oversee coordination of the technical digital creation and the website maintenance team.

In charge of hosting, managing and co-ordinating hackathons on behalf of codecell RGIT.

Papers

This contribution features a novel technique of abstracting dynamic heat-map as the sequential input to custom LSTMs and implements a spinal health index estimator using only an input video of patients performing predetermined tasks.

4th International Conference on Smart Systems and Inventive Technology (ICSSIT 2022)
Preprint + Acceptance
Rudramani Singha Manan Lad Gaurav Shipurkar Anish Rohekar Chandar Chauhan Nilesh Rathod

This paper presents an end-to-end learning-based system that utilizes a multi-branched network architecture to estimate the speed of vehicles from only two time-varied, unprocessed images. And further proposes an optimized method for dataset generation and license detection

4th International Conference on Smart Systems and Inventive Technology (ICSSIT 2022)
Preprint + Acceptance
Rudramani Singha Manan Lad Gaurav Shipurkar Anish Rohekar Chaitanya Thombare Sunil Wankhade

Patents

This patent describes an end-to-end approach to predict the user's intent from a feature-engineered EEG dataset using bidirectional recurrent neural networks.

Application ID: 202121031007
Check Status
Rudramani G. Singha

This patent describes a deep learning approach for an optimized extraction of semi-structured data and the process to cross-validate the data with the live user to authenticate their identity in real-time.

Application ID: 202111043497
Check Status
Rudramani G. Singha Chander Chauhan Arjun Jaggi

Application ID: 202111034856
Check Status
Rudramani G. Singha Manan Lad Chander Chauhan Arjun Jaggi

Application ID: 202111058826
Check Status
Rudramani G. Singha Chander Chauhan Arjun Jaggi

Licenses & certifications

Credential and Verification available at LinkedIn .

Projects

This is the official implementation for one of my research papers

Demo
Keras Tensorflow Blender 3D Deep Learning

This project extracts all the text from a live camera and renders it back on screen to display it to the user

Demo
Tesseract OCR Django GoogleVision API Django

Utilizes the convolutional neural networks to recognize the faces of friends

Demo
Convolutional Neural Networks Keras Tensorflow Fundamentals

Implements YOLO v4 on Raspberry pi with a camera module to intelligently take photo by looking at the gesture of the person in the frame and further sync it to the cloud

Demo
Tensorflow Yolo v4 Raspberry pi 3B TFlite

Skills & Tools

Python R C/C++ Java MATLAB CUDA SQL JavaScript LaTeX Markdown PHP HTML CSS Bootstrap Retriever-Reader Abstractive Summarization Object Detection Image Segmentation Question Answering Blender Autodesk Maya AutoCAD Unreal Engine Unity Photoshop Premiere Pro After Effects PyTorch TensorFlow Scikit-learn Keras Spark JAX Elasticsearch Pandas NumPy Git Flask Django Docker AWS GCP Blender Autodesk Maya AutoCAD Unreal Engine Unity Photoshop Premiere Pro After Effects

Get in Touch

Thank you for taking the time to go through my profile.
Should you need any further information, please do not hesitate to contact me.