Available for opportunities

Pranjal
Gupta

CS undergrad building AI/ML systems — from satellite imagery reconstruction to real-time computer vision and safety tech.

📍 Bangalore / Delhi · B.Tech CSE, University of Delhi · 2024–2028

View Projects Download Resume

About

Who I am

I'm a first-year Computer Science student at the Faculty of Technology, University of Delhi, with a CGPA of 8.20. I spend most of my time at the intersection of machine learning and real-world systems — the kind of work where model performance is measured in decibels or lives kept safer, not benchmark leaderboard positions.

My recent focus has been computer vision: I built a gesture-controlled drone system at Pracverse using MediaPipe and OpenCV, and developed a SAR-guided cloud removal pipeline for ISRO satellite imagery that reconstructs cloud-free scenes at 33.62 dB PSNR — a quantifiable result I'm proud of.

Outside pure ML, I've shipped a cross-platform safety app (Flutter + Twilio), a multi-role healthcare platform with an AI chatbot, and an automated eligibility system with OCR. I like building things that have a clear user on the other end.

Experience

Where I've worked

Computer Vision Intern

Pracverse

Apr 2026 – Aug 2026 · Delhi, India

Designed and built a gesture-controlled drone system — real-time hand-gesture capture translated into flight commands, reducing operator cognitive load significantly over joystick interfaces.
Developed the full hand-tracking and gesture-recognition pipeline using OpenCV and MediaPipe: landmark detection, gesture classification, and command mapping at video framerate.
Worked end-to-end: from data collection and pipeline prototyping to integration with drone control APIs and latency optimization for real-time operation.

Projects

Things I've built

ISRO Hackathon

33.62

dB PSNR

0.96

SSIM

LISS-IV SAR-Guided Cloud Removal

Reconstructing cloud-free satellite scenes by fusing cloudy optical imagery with Sentinel-1 SAR data.

U-Net (ResNet-34 encoder) trained to fuse multi-source satellite inputs — SAR backscatter fills where clouds block optical sensors.
Validated at 33.62 dB PSNR and 0.96 SSIM — strong structural fidelity to ground-truth cloud-free imagery.
Built full preprocessing pipeline and an interactive Streamlit demo for end-to-end cloud removal inference.

Python PyTorch U-Net / ResNet-34 Sentinel-1 SAR Streamlit

GitHub

NaariRakshak

AI-powered women's commute-safety companion — SOS, live tracking, real-time alerts.

Cross-platform mobile app (Flutter/Dart) with one-tap SOS broadcasting via Twilio API.
Live location sharing over OpenStreetMap routing — no proprietary map lock-in.
Background location tracking and trusted contact management with persistent secure storage.

Flutter Dart Twilio API OpenStreetMap

GitHub

Genricycle

DBMS-backed healthcare & sustainability platform — four user roles, one system.

Multi-role platform serving patients, doctors, delivery partners, and labs — telemedicine, lab tracking, pharma logistics in one codebase.
AI chatbot for patient support; secure backend handles transactions, orders, and profile management.
Full-stack: Python/TypeScript backend, relational schema design, REST API layer.

HTML TypeScript JavaScript Python CSS

GitHub

ELIGIFY

Automated exam eligibility checking — dual-strategy PDF parsing with OCR fallback.

Flask app automating eligibility checks via dual-strategy PDF parser: PyPDF2 for structured text + Tesseract OCR fallback for scanned documents.
OpenCV preprocessing pipeline improves OCR accuracy on low-quality document scans.
MVC architecture, REST API, SQL backend — production-ready structure, not a script.

Python Flask OpenCV Tesseract OCR HTML5 / CSS3

GitHub

Education

Academic background

Faculty of Technology, University of Delhi

B.Tech — Computer Science & Engineering

2024 – 2028 CGPA 8.20 / 10
Presidency PU College

Intermediate (Class XII)

2023 92%
R.T.N.E.T Public School

High School (Class X)

2021 GPA 8.9 / 10