I'm a software engineer with experience across robotics, infrastructure, and automation. I most recently worked at Verkada on platform infrastructure, improving Terraform-based IaC workflows and building reusable modules. Before that, I spent over two years at Amazon Fulfillment Technologies & Robotics, where I shipped computer vision enabled systems and scaled device management pipelines across large fleets.
I earned my Master's in Computer Science from The University of Texas at Austin with a 3.96 CGPA and my B.Tech in Computer Science and Engineering from Indian Institute of Technology, Kharagpur, graduating 4th in my department with a 9.62/10 CGPA. My academic journey began in Goa, where I ranked second in the state in the 10th ICSE exams, secured 95.1% in 12th Boards, and earned a 3712 rank in JEE Advanced.
Served as a Teaching Assistant for the course CS 330E Elements of Software Engineering 1. Conducted weekly office hours, graded assignments and tests.
Developed a MiniHDT package in C to send TCP packets to the MCU on the Scout bot for controlling the Actuators and Ultrasonic sensors. This package ran a TCP server to accept connections from multiple client nodes. Created an Actuator service to send TCP messages to MiniHDT to move the Scout forward, in reverse, turn left etc. Implemented code to orchestrate the Ultrasonic service and Actuator service to make the Scout stop on detecting obstacles.
Served as a Teaching Assistant for the course CS 388G Algorithms : Techniques and Theory (MCSO program). Conducted weekly office hours, graded assignments and tests.
Geometric Packing - Squares into Disks: Worked on proving NP-hardness of packing given squares into a disc by a reduction from 3-PARTITION. Determined the dimensions of Framing squares. Created illustrations using the Ipe extensible drawing editor. Introduced the concept of free move to facilitate the proof. Showed the uniqueness of packing the Framing Squares into a unit disc central to the required proof.
Worked on Next Sentence Generation. Compared the performance of pretrained Transformer and pretrained RNN based model with the equivalent models trained from scratch. Worked on Feature Evaluation of Dialog Models. Extracted features such as Dialog Acts, Emotion-Empathetic Intent from utterances in dialogues. Trained a feature dynamics model(LSTM) and compared the performance of Pretrained Transformer, Pretrained RNN based model and DialoGPT model by calculating perplexity for their self play chats.
Outfit Recognizer: Used the IBM Cloud Watson Studio along with its Visual Recognition Service to create an Outfit RecognizerModel which was able to classify Indian clothes into classes such as Kurtas, Sherwanis, Sarees, Lehangas etc and tried to integrate itwith Android Studio. Flower Retailer Chatbot: Used the Watson Assistant service to create a chat bot for a flower retailer which is capable of answeringquestions related to locations and timings of stores and can also recommend flowers for an occasion
Explored the state-of-the-art manufacturing processes for optic fibres, microwave antennas, and base station antennas for 4G cellular communication. Worked on reducing dipole wastage in base station antennas.