Prashish Paudel

Software Engineer · ML / NLP · New York

Prashish Paudel

Software engineer at Blink Voice, working on AI-assisted digital communication at the intersection of voice, language, and people.

Previously at Quantiphi (conversational AI). MS in CS from University of Toledo. Research sits between NLP and healthcare — simulation-based medical assessment and wearables in post-operative cardiac care.

01 — Research

Published work.

  • 01

    Automated Multimodal Performance Evaluation in Simulation-based Medical Education using Natural Language Processing

    ACM/IEE 14th International Conference on Cyber-Physical Systems

    Designed an innovative, integrative solution for simulation‑based medical education performance assessment. Utilized Natural Language Processing(NLP) coupled with precise physiological data analysis for a comprehensive, holistic multimodal evaluation of learner performance at instructor, learner, and environment level.

    Read
  • 02

    Role of Smart Wearable Devices in Postoperative Care of Cardiac Surgery Patients

    59th Annual Meeting of The Society of Thoracic Surgeons

    This study investigates the role of smart wearables in post-surgical cardiac care. Over 60 days, these devices collected vital health data from 20 patients, enabling timely remote interventions. The findings suggest that such technology could potentially cut down on readmissions and serious post-discharge complications.

    Read

02 — Selected Work

Things I’ve built.

  • 01

    Color GAN

    Built a deep neural network model for colorizing grayscale images. Used Pix2Pix GAN a type of Conditional Generative Adversarial Network (cGAN) with a U- Net architecture to train the model on 1100 coastal images dataset and achieved great accuracy on test image dataset.

    PythonTensorflowKerasCNNGANU-Net
    View Source
  • 02

    Sentiment Analyzer

    An application analyzing newspaper sentiment using a Stacked BiLSTM model. Built a News Sentiment Analyzer and deployed a Flask app displaying live sentiment of English News from Nepalese newspapers like Republica, Himalayan Times, and Kathmandu Post.

    PythonKerasLSTMFlaskWord2VecNLP
    View Source
  • 03

    Maritime Vessel Prediction

    Built a model to identify the trajectory of Maritime Vessel from AIS data using Unsupervised clustering technique. Also, analyzed latitude, longitude, speed, and angle to derive a model that predicted the trajectory of a vessel with great accuracy.

    PythonSeabornK-Means ClusteringScikitLearn
    View Source
  • 04

    Neuroplay

    Built a feed forward neural networks to play Flappy bird game on its own. Used NEAT(Neuro Evolution of Augmented Topologies) algorithm to opimize the neural network and accomplished this without using any high level API's and only using NumPy.

    PythonNumPyGenetic AlgorithmNeural Net
    View Source

03 — Toolkit

Things I work with.

Languages

Python, JavaScript, PHP, C, C++, HTML, CSS, VHDL

Frameworks & Libraries

React.js, Node.js, Fat-Free, jQuery, AJAX, NumPy, Pandas

ML / Deep Learning

TensorFlow, Keras, NLP, CNN, LSTM, GAN, Transformer, OpenCV

04 — About

A short history.

Outside of code: trekking (Annapurna Base Camp, Mardi, Gosaikunda, Rara), soccer on the weekends, and a slow but persistent plan to reach Everest Base Camp.

  • 2013Started Computer Engineering Degree at Tribhuvan University, Nepal.
  • 2017Completed Computer Engineering Degree from Tribhuvan University, Nepal.
  • 2018Started working as Software Engineer at FOCUSONE Nepal Pvt.Ltd, Nepal.
  • 2021Started MS in Computer Science at University of Toledo, OH, USA.
  • 2022Worked as Conversational Chatbot Engineer Intern at Quantiphi Inc, MA, USA.
  • 2023Completed MS in Computer Science from University of Toledo,OH, USA
  • 2024Started working as Software Engineer at Blink Voice, NY, USA

05 — Recognition

Certifications & awards.

  1. Deep Learning Specialization — DeepLearning.AI

  2. Graduate Assistantship with Full Scholarship — University of Toledo

  3. Executive Member, UToledo Club for Cyber Security

  4. Poster presentation, ACM/IEEE 14th International Conference on Cyber-Physical Systems