Skip to content
Pratyush Pangotra

Pratyush PangotraOpen to internships

I build electronics that work outside the lab, and the software that makes them useful.

Second-year B.Tech ECE student at MIET Jammu and co-founder & COO of Eusebeia, where we're building SETU, an offline emergency mesh for disaster survivors. I work across ESP32 hardware, Python and data, and AI-assisted product building. I'm looking for internships in product and software, with data and AI as a strong second track.

Studying
B.Tech ECE, MIET Jammu
Building
SETU at Eusebeia
Based in
Jammu, India

Eusebeia

Eusebeia is an early-stage startup I co-founded with two friends. We're small, we build our own hardware and software, and our first product is SETU.

As co-founder and COO, I handle operations and planning, and I work hands-on on how SETU is built and tested.

CEO
Manit Mahajan
CTO
Tanmay Mahajan
COO
Pratyush Pangotra

eusebeia.in(opens in a new tab)

Eusebeia's product

SETU

Survivor Emergency Transmission Unit

“A bridge that doesn't need a network.”

In development

When a disaster takes cell towers down, the people who most need to be found are the ones who can't send anything. SETU turns the phones around them into the network.

How a SETU report travelsA survivor's phone sends a signed report over Bluetooth to a nearby phone, which relays it to another phone, which relays it to the responder command center. No cell tower is involved.No cell networkBluetooth hopBluetooth hopBluetooth hopSurvivorspeaks, AI triagesRelay phoneforwards, can't editRelay phoneforwards, can't editCommand centerverifies, maps, dispatches
A report is signed on the survivor's phone, so it arrives exactly as it was sent.

How it works

  1. Speak

    A trapped person records a short spoken report. No signal or internet is needed.

  2. Triage on the phone

    On-device AI transcribes the report, gives it a START triage tag, and packs it into a tiny payload signed with Ed25519.

  3. Hop

    Nearby phones running SETU pass the payload along over Bluetooth, one hop at a time. Because it's signed, any change a relay makes is detected.

  4. Map and dispatch

    At the command center, the dashboard verifies each report, clusters related ones, maps them, and helps responders decide who to send where.

Design choices

Ed25519-signed payloads
Relays carry reports but can't tamper with them.
START triage tags
Every report arrives already sorted by urgency, using the standard mass-casualty triage scheme.
Automatic clustering
Reports about the same incident are grouped, so responders see incidents rather than a flood of messages.
Offline multilingual translation
A survivor and a responder don't need to share a language, or an internet connection.
Zero-backhaul command center
The responder dashboard runs without any link to the outside internet.

Projects

Featured hardware project

ForestGuard

IoT wildfire detection network

A real-time forest fire detection network built on ESP32 microcontrollers. Each node fuses readings from a DHT22 temperature sensor and an MQ-135 air quality module to detect fire signatures. Alerts go out through a Telegram bot in Hinglish, with auto-normalisation, so forest rangers can act on them without technical training. Tested with incense smoke.

  • ESP32
  • Embedded C
  • DHT22
  • MQ-135
  • Telegram Bot API

View on GitHub(opens in a new tab)

Controller
ESP32
Sensors
DHT22 + MQ-135 (sensor fusion)
Alerts
Telegram bot, in Hinglish
Tested with
Incense smoke
Competition
YESIST12 prelims, April 2025
My role
Team lead, Team Pratham
Aligned with
UN SDG 15, Life on Land
  • Smart School Bell Automation

    Hardware

    Automated school bell scheduling built on ESP32 boards that talk to each other over ESP-NOW, so the bell units don't depend on a Wi-Fi router.

    • ESP32
    • ESP-NOW
    • C++

    View on GitHub(opens in a new tab)

  • ESP32-S3 Macropad

    Hardware

    Custom productivity macropad built on an ESP32-S3 N16R8 (16 MB flash, 8 MB OPI PSRAM) with an SSD1306 OLED. Four programmable keys for VS Code, Claude, YouTube and Terminal: StreamDeck-style shortcuts with no subscription.

    • ESP32-S3
    • SSD1306 OLED
    • USB HID

Competitions

  • April 2025YESIST12 prelims, ForestGuard, as team lead of Team Pratham

Research

Research in progress. Not yet published.

Non-invasive blood group prediction from fingerprints

My own ongoing research into whether a fingerprint image carries enough signal to predict ABO/Rh blood group, as a possible needle-free screening aid. I'm using transfer learning on a pretrained ResNet18 and treating it as an 8-class image classification problem.

Model
ResNet18, transfer learning
Task
8-class blood group classification
Test accuracy so far
91.2%

These are early results, and I'm still refining the method.

Skills

Embedded & IoT

Where most of my hands-on work is.

  • ESP32 / ESP32-S3
  • Arduino
  • nRF
  • ESP-NOW
  • MQTT
  • LoRa
  • I²C / SPI / UART
  • Sensor integration
  • Embedded C

Python & data

For analysis, visualisation and my ML research.

  • Python
  • pandas
  • NumPy
  • matplotlib
  • MySQL
  • C

Web & product

I build web products with AI-assisted workflows. My strength is scoping, product decisions and shipping, not deep front-end specialism.

  • Next.js
  • React
  • Supabase
  • Flask

Tools

Day-to-day.

  • Git & GitHub
  • VS Code
  • PlatformIO
  • Arduino IDE
  • Telegram Bot API
  • Vercel

Education

  1. 2025 – present

    B.Tech, Electronics & Communication Engineering

    MIET (Autonomous), Jammu, 2nd year, Semester III

  2. Class XII

    Senior Secondary, Science

    CBSE, 82%

  3. Class X

    Secondary School

    CBSE, 87.4%

Contact

I'm looking for internships in product, software, and data/AI, and I'm happy to talk about SETU or any of the hardware here. Email is the fastest way to reach me.

pratyushpangotra74@gmail.com
Based in
Jammu, India