Python
Django
FastAPI
Gradio
Gurobi
Huggingface
Langchain
Numpy
OpenCV
Pytorch
Ray
Scikit Learn
Tensorflow
Tornado
Computer Vision With Audio Aid
Specifically designed for blind people. Gives a description of detectable things in a given image/frame.
Specifically designed for blind people. Gives a description of detectable things in a given image/frame.
Stock Maket Trading Bot
Developed a stock trading system using the 5paisa Python API and backtesting strategies with Python Backtrader, enabling automated and data‑driven trading decisions
Developed a stock trading system using the 5paisa Python API and backtesting strategies with Python Backtrader, enabling automated and data‑driven trading decisions
Human Activity Recognition
Evaluated various deep learning models (CNN, CNN‑LSTM, LSTM‑CNN, LSTM, CNN‑GRU,GRU‑CNN, GRU), including convolutional, gated recurrent, and long short‑term memory networks, for accurate human activity recognition
Evaluated various deep learning models (CNN, CNN‑LSTM, LSTM‑CNN, LSTM, CNN‑GRU,GRU‑CNN, GRU), including convolutional, gated recurrent, and long short‑term memory networks, for accurate human activity recognition
Gait classification framework based on Ensemble Learning
Ensemble learning on the task of gait recognition using CNN
Ensemble learning on the task of gait recognition using CNN
Audio Denoiser
Engineered an innovative real-time audio denoising system to effectively filter out ambient noise
Engineered an innovative real-time audio denoising system to effectively filter out ambient noise
Manufacturing Scheduler
Streamlined Production Enhancement: Optimizing and Scaling Scheduler for Peak Efficiency Elevating Genetic Algorithm Scheduler: Fine-Tuning Performance across Multiple VMs Empowering Pharma Production: Enabling Medicine Scheduling for a Prominent Pharmaceutical Enterprise
Streamlined Production Enhancement: Optimizing and Scaling Scheduler for Peak Efficiency Elevating Genetic Algorithm Scheduler: Fine-Tuning Performance across Multiple VMs Empowering Pharma Production: Enabling Medicine Scheduling for a Prominent Pharmaceutical Enterprise
Card Detection
Developed an OpenCV-powered application to implement advanced feature recognition tailored for precise metro card identification. Leveraged PaddleOCR to seamlessly extract text content from the identified metro cards. Streamlined the card scanning process and automated fund replenishment, optimizing user experience and efficiency
Developed an OpenCV-powered application to implement advanced feature recognition tailored for precise metro card identification. Leveraged PaddleOCR to seamlessly extract text content from the identified metro cards. Streamlined the card scanning process and automated fund replenishment, optimizing user experience and efficiency
Graph Neural Network Based Delay Estimation of Large Computer Networks
Utilized graph neural network (GNN) based route estimation using routenet to analyze the impact of node and path interactions on network performance metrics, enabling efficient network design and optimization.
Utilized graph neural network (GNN) based route estimation using routenet to analyze the impact of node and path interactions on network performance metrics, enabling efficient network design and optimization.
Packaging Scheduler
Developed a solver application for multi‑day, multi‑shift pharmaceutical packaging management, optimizing production efficiency and reducing costs by upto 80%
Developed a solver application for multi‑day, multi‑shift pharmaceutical packaging management, optimizing production efficiency and reducing costs by upto 80%
Rapid Investigator
Led the full‑stack development of a data pipeline with a user‑friendly interface, enabling data‑driven insights and improved decision‑making in pharmaceutical operations.
Led the full‑stack development of a data pipeline with a user‑friendly interface, enabling data‑driven insights and improved decision‑making in pharmaceutical operations.
Medical Platform
Orchestrated and scaled a medical platform, integrating OpenEMR and Magento using Kubernetes, improving reliability, scalability, and performance
Orchestrated and scaled a medical platform, integrating OpenEMR and Magento using Kubernetes, improving reliability, scalability, and performance
User Pipeline
Developed a user AI agent that automates user modifications to AI output, enhancing user experience and streamlining workflows.Conducted thorough analysis of user preferences and modifications to enhance optimal outcomes Seamlessly integrated the workflow onto cloud platforms for enhanced accessibility and collaboration
Developed a user AI agent that automates user modifications to AI output, enhancing user experience and streamlining workflows.Conducted thorough analysis of user preferences and modifications to enhance optimal outcomes Seamlessly integrated the workflow onto cloud platforms for enhanced accessibility and collaboration
Head Pose Detection
Designed and developed a sophisticated model for accurately detecting the pose of a person’s head. Enhanced the model’s predictive capabilities to achieve significant improvements in accuracy and performance
Designed and developed a sophisticated model for accurately detecting the pose of a person’s head. Enhanced the model’s predictive capabilities to achieve significant improvements in accuracy and performance
Optimized Training
Engineered a Robust Training Pipeline to Evaluate User Models Across Optimizers, Loss Functions, and Data Selection
Engineered a Robust Training Pipeline to Evaluate User Models Across Optimizers, Loss Functions, and Data Selection
MLOps Setup
Established a Kubeflow-based infrastructure to streamline ML team workflows. Developed tailored Docker images to cater to specific requirements of the ML team. Implemented seamless local and remote storage mapping via cloud integration. Strategically partitioned Nvidia A100 GPUs, fostering efficient resource sharing among cross-functional team members
Established a Kubeflow-based infrastructure to streamline ML team workflows. Developed tailored Docker images to cater to specific requirements of the ML team. Implemented seamless local and remote storage mapping via cloud integration. Strategically partitioned Nvidia A100 GPUs, fostering efficient resource sharing among cross-functional team members
Image Repainting
Developed a sophisticated repainting solution where one can select an object through prompts (click, text, or mask) and repaint it to the desired color without any shadow or lighting issues.
Developed a sophisticated repainting solution where one can select an object through prompts (click, text, or mask) and repaint it to the desired color without any shadow or lighting issues.