Samanti Das

Deep Learning for Computer Vision, Image/Point Cloud Processing, Object Detection and Tracking, Autonomous Systems.

ENSNARE Project (EU-funded) Domain: Data/Image Processing · BIM · Robotics 1. Developed algorithms to maximize solar panel placement on building facades. 2. Built a FreeCAD plugin for automated 3D reconstruction from 2D images. 3. Implemented FreeCAD–ROS real-time communication for robotic assembly workflows. 4. Tutored Automation and Robotics in Building Renovation. 5. Contributed to 3 conference papers and 1 journal article.

Chair: Electronic Design and Automation Role: Development of ETISS ISA based on RISC-V developed by the Infineon Technologies Tech stack: C, C++, Python, GitHub Actions (for continuous integration)

Title: Improving the robustness of lightweight and deep semantic segmentation networks. Role: Development of an auto-augmentation strategy that jointly optimizes an augmentor network and the semantic segmentation network. Application: Safety in autonomous driving and medical diagnosis Tech stack: Python, PyTorch

Lab: Object Recognition Lab under Dr. Hamidraza Houshiyar Role: Automatically align point cloud from LiDAR to the ground plane by fusing information from an IMU sensor. Tech stack: Python, C++, ROS

Roles: 1. Literature survey of SOTA grasping algorithms. Focus on removing objects from clutter using RGB-D sensor and Franka Panda arm. 2. Comparitive study between two grasping algorithms: NVIDIA Graspnet and Volumetric Grasping Network. 3. Simulating grasping using PyBullet and in real. Tech stack: Python, Tensorflow, Keras, C++, ROS

Title: Targeting Paris Convention in the scope of COVID-19, as part of the course ‘Applied Machine Intelligence’. Roles: 1. Predicting carbon dioxide emissions from 14 European countries via time-series ML models (LSTM, GRU, CNN, MLP) until 2030 2. Building a user friendly web application to predict the outcome taking into account the lockdown measures that were applied during the pandemic Tech stack: Python, Numpy, Pandas, Keras, Scikit-Learn, TensorFlow, Machine Learning, Neural Networks etc.

I was a Trainee who was asked to design the prototype of a Temperature Monitoring System and a Frequency Controller of Induction Oven to be used in ECR Ion Source

Thesis: Chair of Media Technology under Dr.-Ing Prof Steinbach Title: Improving the robustness of lightweight and deep semantic segmentation networks. Role: Development of an auto-augmentation strategy that jointly optimizes an augmentor network and the semantic segmentation network. Application: Safety in autonomous driving and medical diagnosis Tech stack: Python, PyTorch

Focus: Computational Photonics, Electronics, VLSI, Data Structures and Algorithms Bachelor Thesis: Analysis of EM wave propagation in photonic components and structures using Finite Difference Techniques. Focus on Space Division Multiplexing and development of passive optical devices. + 1 journal paper: https://doi.org/10.1007/978-981-15-2926-9_13

Photonics, Optoelectronics, Computer Architecture, Digital Logic Design, VLSI