Developer Advocate Engineer, Dexmate

Pragya Sen

Hii, Pragya here! I'm currently working at Dexmate as their Developer Advocate Engineer. For those who don't know, Dexmate is building the next generation of physical AI ✨

While I have a strong technical background with a Master's in Data Analytics and a Bachelor's in Electronics, I also really enjoy connecting with new people and creating content (check out my blog). I have always wanted to put my technical skills to use in a way that touches many lives and have found a way to do so with this role at Dexmate :)

Prior to moving to the US, I worked at Sandvine Technologies as a Software Test Engineer for 1.5yrs.

I'm using this page to document all my technical and creative pursuits. Feel free to look around!

Email

pragyas@bu.edu pragya.sen1@gmail.com

Location San Jose, CA

Follow

Sept 2024 – Jan 2026

Master's Degree

Boston University - Applied Data Analytics

Aug 2019 – June 2023

Bachelor's Degree

R.V College of Engineering - Electronics and Communication Engineering

Aug 2026 – Present

Developer Advocate Engineer

Dexmate

I'm currently working at the intersection of AI/ML, robotics and developer experience. I support developers using Dexmate's Physical AI platform and work closely with the engineering, product and marketing teams. Responsibilities include developing sample applications, technical resources, speaking at AI/Robotics conferences and connecting with developers to grow the developer ecosystem.

March 2026 – Aug 2026

AI Engineer

Handshake

Worked on model training and hallucination reduction in LLMs with a focus on prompt engineering. Documented evaluation insights, failure patterns and edge cases to drive continuous improvements, specifically on NVIDIA's Nemotron-Nano-30B-A3B-Omni.

Jan 2026 – April 2026

Web Development Intern

Casa Carlini

Analyzed website traffic and reader behavior using analytics and performance metrics for improvements in usability, page layout and site performance. Led a WordPress to Shopify migration alongside backend database operations using SQL/MySQL.

March 2023 – Aug 2024

Test Engineer-II, Test Engineer-I, Software Intern

Sandvine Technologies

Worked on Software test automation using Python, Selenium, Appium, Flask, etc. Developed a test-suite for automated testing of applications for traffic generation and provided multi-site support (4 sites: Bangalore, Waterloo, Dallas, Malmo). Handled the Bangalore lab-site as the primary owner. It had two setups (Fixed and Wireless) with a total of more than 24 devices. More than 1000 test cases per setup were run on a daily basis.

Projects

look at what I've built

LLM Inference Benchmarking

Developed a toolkit for benchmarking local LLMs and visualizing the results in an interactive Streamlit dashboard. It measures key inference metrics: TTFT (Time to first token), End-to-end Latency, Tokens generated, Tokens per second.

View on GitHub →

Journaling Companion

Built a journaling companion using Llama via Groq with guided conversational prompts, real-time sentiment analysis (CardiffNLP RoBERTa), theme detection (BART-MNLI) and voice-to-text journaling (Whisper Large v3). Privacy-first architecture with SQLite for daily conversation storage and Google Drive OAuth integration.

View on GitHub →

Patient Voice Agent

Developed an automated voice bot that calls a test line, simulates realistic patient scenarios, records conversations and saves transcripts for bug reporting.

View on GitHub →

RAG-based Author Metadata Extraction

Developed a system that extracts affiliation and contact details of the first and last authors from the enclosed PDF file containing academic paper references and outputs the data into structured Excel files. Leveraged LLMs to interpret complex text and add online search capabilities to cross-reference and validate data against academic databases and institutional websites, ensuring high accuracy.

View on GitHub →

Blur, Noise & Saturation Handling in Images

Developed a model for Deblurring and Denoising images with outliers (saturation) using the help of an optimum GAN module. Implementation done using Python.

View on GitHub →

Large-Scale Face Matching

A comprehensive approach to address various real-world challenges in face matching within face search systems. The implementation has been done using Python.

View on GitHub →

Data Warehousing

Developed an efficient Data-warehouse structure from 2 large datasets (Cosmetics and Consumer-products) from the official US Govt. website to answer various business questions. PostGre SQL has been used for implementation.

View on GitHub →

Sentiment Analysis

Performed sentiment analysis on restaurant reviews using two different approaches: a traditional machine learning model (Naive Bayes) and a modern deep learning model (BERT). Provided an interactive visualization using Gradio for end-user interpretation.

View on GitHub →

Customer Segmentation & Order Cancellation Analysis

This project deals with Customer Segmentation based on purchasing power & Order Cancellation analysis of Amazon Sales data. The implementation is done using Python.

View on GitHub →

Dental Cavity Detection

The project has employed visual pictures of teeth and applied a deep convolutional neural network (CNN) to categorize the teeth into caries or non-caries. Implementation done using Python.

View on GitHub →

Cardiac Dynamics

Worked on the mathematical modelling of Cardiac function of the heart when the heartbeat is too fast (tachycardia) or too slow (bradycardia). Visualization done using MATLAB.

View on GitHub →

Red-Wine Analysis

The project aims to determine which chemical features are the best indicators of red wine quality and use the most appropriate method to model the same. Has been implemented using R programming.

View on GitHub →