Avigyan Sinha

About Avigyan Sinha

An AI Researcher and Developer with over a decade of experience, focused on Computer Vision, NLP, Multi-modal analysis, Explainable AI (XAI), and Generative AI. Throughout my career, I have led multiple research teams, customized and implemented cutting-edge deep learning architectures like Convolutional Neural Networks, Long Short-Term Memory, Gated Recurrent Unit, AlexNet, VGGNet, GoogLeNet (Inception), Residual Network, SqueezeNet, MobileNet, Faster R-CNN, Mask R-CNN, Single Shot Detector, YOLOv11, DenseNet, UNet, Autoencoder, Generative Adversarial Networks, Transformers, Attention Mechanism, DDPM, DDIM from scratch in projects, and contributed to the development of innovative AI models. My unique strength lies in applying the power of computer vision to diverse fields such as disease detection, and even autonomous drone identification. I\'ve delivered projects ranging from automatic brain hemorrhage detection using neural networks to real-time drowsiness detection via video processing—all leveraging the power of deep learning. My experience is not just theoretical, it\'s backed by tangible results. Projects under my guidance have achieved impressive accuracies: language identification from speech signals at 98%, breast cancer detection at 99.9%, and hand gesture recognition at 96%, to name a few. My work on drone detection technology has even resulted in a patent. Being a part of leading MNCs like Vehere and working on NASA (and ISRO)-funded projects has equipped me with the ability to deliver under stipulated time frame while keeping innovation at the forefront. Looking forward to roles that transform complex problems into applicable machine learning solutions. Whether it\'s developing cutting-edge algorithms or guiding a team through a challenging project, I\'m committed to utilizing ML techniques for product development, to build commercially viable, cost-effective, optimal technological solutions. My interest aligns with industrial R&D positions in the fields of Deep Learning, Computer Vision, NLP. I believe that machine learning holds the key to improve the quality of lives, and extend our imagination, capability beyond physical impediments.Skills: Python, C++, C, MATLAB, Java, Keras, Tensorflow, PyTorch, OpenCV, MXNet, Google Colab, Anaconda, Jupyter Notebook, SHAP, Pandas, Scikit-image, Scikit-learn, PyQt, XGBoost, Scipy, PIL, NLTK, NumPy, CuPy, ImageJ.Email: a••••••••@gmail.com, a•••••••@alumni.jh.eduGitHub: https://github.com/avigyanGoogle Scholar: https://tinyurl.com/avigyansinhaWebsite: https://avigyan.github.io

Experience

1. Senior Engineer i, R & d (Deep Learning, Computer Vision)

   Vasoptic Medical, Inc

   Apr 2023 — Present · Columbia, MD, US

   Led multiple research teams in developing novel classification, image segmentation algorithms and optimizations, resulting in improved predictive accuracy. • Managed the creation, launch, and delivery of multiple new image analytics and multi-modal imaging software products to customers.• Developed novel blood vessel segmentation, optic disc segmentation, and artery-vein classification algorithms for Laser Speckle Contrast Imaging (LSCI) video data.• Conducted analysis of LSCI video data for eye health, contributing to AI model development for improving diagnosis rates across different ocular conditions in a diverse cohort of patients.

Education

• Indian Institute of Technology, Kharagpur

  Bachelor of Technology (Honours), Electronics and Electrical Communication Engineering

  2009 — 2013

• The Johns Hopkins University

  Master of Science in Engineering, Electrical and Computer Engineering

  2013 — 2015

Skills

• C
• Programming
• Optimization
• Xilinx
• Java
• Object Tracking
• Matlab
• Computational Modeling
• Visual Studio
• Signal Processing
• Pattern Recognition
• Computer Vision
• Algorithms
• Image Processing
• C++
• Xilinx Ise
• Technical Computation, Simulation and Programming
• Linear Algebra
• Machine Learning
• Research
• Physics
• Linux
• Electronics
• Opencv
• Statistical Data Analysis
• Cadence
• Verilog