I’m Sai Dharshan!

CSE @ Sathyabama Institute of Science & Technology · ML & Defence Tech

I’m a Computer Science Engineering undergraduate in Chennai, working where machine learning meets defence technology. I lead team GOODFELLASS on a terrain-aware counter-drone system designed to hear drones past mountain ridges, I’m developing a drone survey system that verifies how much legacy waste actually leaves city dumpsites, and I’m founding Aeroscan Defence Systems to take that counter-drone work to high-altitude regions.

View Projects Get in touch

Contact

saidharshan922@gmail.com github.com/Sai-Dharshan linkedin.com/in/saidharshan

Skills & software

Toolkit

01Languages

Python
Java
C++

02ML & Vision

Machine learning
PyTorch
Segmentation (U-Net, SegFormer)
Monte Carlo uncertainty
Edge AI (NVIDIA Jetson)

03Drones & Geospatial

Photogrammetry (WebODM)
Point clouds (CloudCompare, PDAL)
DEM change detection
Thermal imaging
rasterio · NumPy
Leaflet web maps

04Hardware & Signals

Acoustic arrays (MEMS, I²S)
ESP32-S3 firmware
Signal processing (GCC-PHAT)
RS-422 signalling
SDR (RTL-SDR, GNSS-SDR)
Ruggedisation (MIL-STD-810H)

Selected work

Projects

Defence sensing and machine learning, built for conditions where the usual tools fail.

Smart India Hackathon 2026 · SIH26050 · DRDO · Team leader, GOODFELLASS

Terrain-Aware Counter-UAS System

The problem

At 4,500–5,000 m, ridges block line of sight and fog and snow blind optics. Hardening the hardware can’t fix that, so the system senses at wavelengths terrain and weather don’t block.

Acoustic arraysSignal processingESP32-S3SDRRuggedisationHardware

What we built

A full-scale 8-microphone nested acoustic array, designed to hear rotors past a ridge; built and tested at sea level
A satellite forward-scatter corridor gate that flags drones crossing mountain passes (designed)
Fixes for two altitude errors: a sound-speed bearing bias of up to ~12° and a ×1.29 rotor-tone shift

GitHub View more

CiiA 2026 · Reworked · In progress

Drone-Based Landfill Survey System

The problem

Cities pay contractors per tonne to clear legacy dumpsites, but drone surveys give one volume number with no error range, and can’t tell waste that left the site from waste only moved around inside it.

What I’m building

Removed tonnage with a 95% confidence interval, from DEM-of-difference change detection and Monte Carlo density modelling
Material-zone segmentation that separates real removal from relocation and fresh dumping
Thermal hotspot tracking with in-flight alerts from an onboard Jetson

DronesPhotogrammetryComputer visionThermal imagingPython

GitHub View more

Founder · Incubation proposal

AeroscanDefence Systems

A defence-tech startup building anti-drone systems for high-altitude regions. Its first and only product is the terrain-aware counter-UAS system above: acoustic and satellite forward-scatter sensing for terrain where line of sight and visibility fail. Aeroscan Defence Systems Pvt. Ltd. is being proposed to the startup incubation cell at Sathyabama Institute of Science & Technology.

Talk about Aeroscan

Focus: Counter-drone systems
Built for: High-altitude regions
First product: Terrain-aware C-UAS
Stage: Incubation proposal
Role: Founder

Events

22OCT 20269:30 AM – 1:15 PM

Organiser

Adversarial Arena

AI attack / defend technical event

Participants attack and defend AI systems across rounds: a hands-on look at adversarial machine learning and AI security, held at Sathyabama Institute of Science & Technology.

Education

Undergraduate

B.E. Computer Science Engineering

Sathyabama Institute of Science & Technology, Chennai
School

Senior secondary schooling

Vani Vidyalaya Senior Secondary & Junior College

Achievements

Smart India Hackathon
SIH 2026 idea submission for SIH26050 (DRDO), as team leader of GOODFELLASS

Let’s worktogether

Building something in ML, sensing or defence tech, or want to collaborate on Aeroscan? Send a note and I’ll get back to you.

Email me