Neeha Chowdary Sajja
OBJECTIVE
I’m a Software Developer with a Master’s degree in Robotics and AI from La Sapienza University of Rome.
Currently, I work in telecom. This site is a collection of my work from my university projects to personal code I write today. Whether it’s detecting objects in the Waymo dataset or working on telecom systems, I focus on building software that is fast and reliable.
BIO.
Education Qualification
High School Study from Central Board of Secondary Education from Jubilee Hills Public School.
Graduated in Bachelor of technology in computer science, from Jawaharlal technological university and
Masters in Robotics and Artificial Intelligence, from Sapienza University of Rome, Italy
Communication Languages
English IELTS score-7
Italian score B2
Hindi (national language)
Telugu ( mother tongue)
Computer Languages
C, C++
javascript
Python
Matlab
VREP
ROS/ROS2
Tensor flow
Proficiency in using productive software and electric and communication skills.
Master Thesis
I have performed my master thesis at the La Sapienza University of Rome, in the topic of cable-suspended parallel robots.
My research was mostly focussed on modelling and kinematic control of the robot.
using this semi-automated Robot many companies can increase their efficiency in their production, As the task of being a crane operator is difficult and the necessity of the transportation of goods is high which can be full-filled by a cable suspended robot.
Here is my Github link with all the original source codes for some more projects aswell.
https://github.com/sapienza123?tab=repositories
MEDICAL ROBOTICS
INTEGRATION OF NEW ENDOSCOPIC INSTRUMENT IN THE DA-VINCI VREP SIMULATOR
The Aim of the Project is to integrate a new endoscopic instrument(Hand) in the Da Vinci Kit in V-
REP and Perform Variety of simulations using the new integrated instrument. Simulations
performed, are done in Virtual environment (V-REP) and Simulated Using the ROS (robot
operating system) as interface to simulate Robot(DVRK) in V-REP, that has the new integrated
instrument(Hand). These Simulations, include Opening and Closing of the new Endoscopic
Instrument.
VISION AND PERCEPTION
Mask_RCNN and LASTM video classification
Computer Vision have gotten to the level, where a lot people are researching on various ways to improve the dash cam on an autonomous vehicle or the vision of the robot and also for improving the imaging application for the radiography and MRI.
With the applications, like Mask_RCNN and other deep learning application designed to improve vision the application we might have chance of improving the way the machines will learn how to view and interact with world around them.
LEARNING IN AUTONOMOUS SYSTEMS
PING - PONG
A ping pong game is where there are two paddles and a ball in between, have to try and get ball pass between the paddles, this game has been build from scratch and trains a Neural Network to get it better over time.
Deep Q Network is built to read it in pixel data from the game of pong .where Deep mind is to read in the pixel data and the score, these two are its inputs to the DQ Network, based on those two it gets better over time, reinforcement learning.
NATURAL LANGUAGE PROCESSING
CHATBOT (JOHN)
In the present world of machine learning and AI there are many different kinds of approaches, for the increase of human comfort with the positive increase in technology, Chatbots which is a conversational technology to increase customer service in many fields. The chatbot here uses neural networks and deep learning so it is slightly intelligent. It is a chatbot to answer trivia questions like for a specific purpose. A deep learning chatbot learns right from scratch through a process called “Deep Learning.” In this process, the chatbot is created using machine learning algorithms. A deep learning chatbot learns everything from its data and human-to-human dialogue. The chatbot here is called JOHN, and he answers to some questions to students about the diag school.
HUMAN ROBOT INTERACTION
EMOTION RECOGNITION
Emotion Recognition is a project where the code is used to know whether the person is sad, happy, angry and any others. In here Web camera of the computer is used for identifying the expressions in Real-time. This is a three step process, the XML file detecting the presence of faces and then we retrain the network with the images on five different categories, Then the importing of the label_image.py program and set up everything in Realtime.
MOBILE APPLICATION AND CLOUD COMPUTING
AN APP FOR BOOK EXCHANGE
In the Online Book Exchange System, there can be 2 possible users; they are the owners and the borrowers of the book. The owners have the capability to post all the details of the book and with their contact information on the other hand borrower who is interested with the listed book will also update their contact details. So the interest of the user matches with any of the book listed with the system. They can directly contact and exchange the book themselves through this system. The purpose of the Online Book Exchange System is to provide quality services. It aims to provide transparency in the borrowing process. The system provides a feature that can easily manage user’s information, book details and manage borrowing transactions.
NEURAL NETWORKS
A Functional Link Adaptive Filter algorithm is introduced to design a map signal into high dimensional features space, we implement two variations of FLAF, namely CFLAF and PBCLAF in three different expansion type. Which are Chebyshev Polynomial expansion, Legendre polynomial expansion and trigonometric polynomial expansion type
INTERACTIVE GRAPHICS
SHADOW MAPPING
The faithful rendering of fully dynamic, hard shadows is a feature of great importance in interactive 3D graphics. Shadows help the understanding of the structure of scenes and contribute strongly to the perceived reality.
the project is a kind of small animated theme example, that is moving of two objects(cuboids) in a room.
The project includes Lights and Textures,
Animations, which includes user interaction like these keys (W key, S key, A key D key) are assigned functions for zoom in zoom out move left move right
Left arrow rotation and right arrow for right rotation alike.
OBSTACLE AVOIDANE
In this project, I focused on the full lifecycle of robotic development—from physical modeling to algorithmic control. Using Webots, I constructed a differential drive robot and the heart of the project is a C++ backend that implements a 'Sense-Think-Act' cycle. By analyzing IR/Lidar data streams, the controller identifies obstacles in the robot’s local costmap and dynamically recalculates the angular velocity to perform a 'seek-and-avoid' turn. This demonstrates a deep understanding of robot kinematics and C++ memory management in a simulation-heavy environmen
OBJECT DETECTION ON WAYMO OPEN DATASET
Working with the Waymo Open Dataset presented a unique challenge due to its sheer scale and the raw format of its sensor data. I performed reverse engineering on the timestamped image sets to 're-stitch' them into fluid video sequences. This allowed me to perform object detection and classification not just on individual frames, but across time providing the model with the temporal context necessary for autonomous navigation tasks. The entire pipeline was built in Python, focusing on efficient data handling and high-precision classification .
Get in Touch
Interested in learning more about me, my work or how we can collaborate on an upcoming project? Feel free to reach out anytime, I would be more than happy to chat.