Chennai, India — Portfolio ’26

Open to internships & collaborations

Darshan C S — Full Stack Developer & AI Engineer

I build products that actually ship.

Pre-final-year AI & Data Science student shipping end-to-end products: React front ends, Django & WebSocket backends, Postgres data layers and ML/LLM brains for healthcare, emergency response, networks and ed-tech.

See the work Let’s talk

Portrait of Darshan C S in a maroon blazer — Darshan C SFull Stack × AI

4 hackathon wins 8.6 CGPA · REC ’28 1 paper under review Scroll to explore

Winner — Smart Ability Hackathon 2026✦ 2× Winner — SIH Internal Round✦ Winner — Kuttykoncepts Hackathon 2026✦ Finalist — CIH ’26✦ FH Aachen — 4.75 / 5✦ AIFB 2026 — IIT Dharwad✦ Winner — Smart Ability Hackathon 2026✦ 2× Winner — SIH Internal Round✦ Winner — Kuttykoncepts Hackathon 2026✦ Finalist — CIH ’26✦ FH Aachen — 4.75 / 5✦ AIFB 2026 — IIT Dharwad✦

React/TypeScript/Django REST/WebSockets/PostgreSQL/Supabase/scikit-learn/LLMs/Three.js/Docker/FHIR R4/ React/TypeScript/Django REST/WebSockets/PostgreSQL/Supabase/scikit-learn/LLMs/Three.js/Docker/FHIR R4/

01 About The person behind the pull requests

Where intelligence meets impact.

Darshan in a maroon blazer walking down a cobbled street

Darshan leaning on a stone balustrade beside an old lantern

Darshan in an orange hoodie in front of a giant Mercedes-Benz star

Darshan in a denim shirt in front of a gothic cathedral portal

Darshan in black standing beside a bronze panther fountain

Darshan leaning against an ornate white facade

Darshan in a white shirt leaning on a silver car at night

I’m Darshan, a pre-final-year AI & Data Science student at Rajalakshmi Engineering College, and an engineer who turns ambitious ideas into products people rely on.

I’m a full-stack developer with an AI engineer’s instincts. I design React and TypeScript interfaces, build Django and WebSocket backends, shape data in PostgreSQL and wire in the machine-learning and LLM systems that make a product think. I care as much about the architecture nobody sees as the details everybody feels.

I lead from the front and I’m comfortable being the first voice in the room: pitching in 24-hour hackathons, running workshops single-handed, hosting conversations with academics and founders. Deadlines make me calmer, not louder, and I don’t stop at “it works”. I stop when it ships.

Technical depth: Pixels to model weights.
Leadership: First to step up.
Confidence: Composed in any room.
Work ethic: I finish what I start.

2025 — 2026

6 hackathons.

SIH ’25 · Smart Ability ’26 · Kuttykoncepts ’26 · CIH ’26 · Adobe ’26 · SIH ’26
and of those…

4 wins.

Smart Ability Hackathon ’26 · Kuttykoncepts Hackathon ’26 · SIH internal round, twice
Adobe University Hackathon 2026

3,00,000+ rivals.

Cleared two of three rounds, nationwide.
June 2026

3 internships.
At once.

Kuttykoncepts · FH Aachen · Gocourse, alongside full-time coursework
International summer internship

95% at FH Aachen.

4.75 / 5 · CyberSol International Summer Internship · MASCOR Institute
Research

1 paper under review.

“Ghost Evidence” · AIFB 2026 · IIT Dharwad
In production

1 platform going live
for the Govt. of India.

EchoSense AI · being deployed at NIEPMD
Deployed in my department

CLICK, where first-years
learn C.

C programming learning + practice platform · built for the 1st-year students of my department at REC
And honestly?

I’m just getting warmed up.

02 Experience Where the work gets forged

Five roles. One relentless year.

2026 — PresentChennai

Full Stack Developer & AI Engineer

NIEPMD National Institute for Empowerment of Persons with Multiple Disabilities

Developing EchoSense AI, an end-to-end tinnitus assessment, sound-therapy and doctor-consultation platform (Winner, Smart Ability Hackathon 2026), now being deployed for NIEPMD.
React + TypeScript front end with a Three.js 3D ear viewer, a clinical-convention audiogram, Web Audio sound therapy and i18next.
Django REST backend with JWT and HL7 FHIR R4, plus a scikit-learn/SciPy prediction engine across seven modules. Playwright E2E, deployed on Render.

React
TypeScript
Three.js
Django REST
FHIR R4
scikit-learn

Jun 2026 — Presentvia REC

Generative AI & Prompt Engineering Intern

Kuttykoncepts

Produces AI-generated study material end to end: course concepts become scripts, AI avatars and fully AI-generated instructional videos.
Designs and iterates prompts so the output stays accurate, curriculum-aligned and consistent across subjects.

GenAI
Prompt engineering
AI video

Jun 2026MASCOR · FH Aachen

CyberSol International Summer Intern

MASCOR Institute, FH Aachen University of Applied Sciences

Four-week immersion and internship in Deep Reinforcement Learning, AI and Neural Networks: two weeks online, two weeks on-site.
Graded 4.75 / 5 (95%) for a 60-hour, 4-credit workload covering Deep RL, Sim-to-Real and Cognitive Robotics.

Deep RL
Neural networks
Sim-to-Real
Robotics

2026 — PresentRemote

SEO Intern

Gocourse Pvt Ltd

On-page SEO for the Gocourse books website: keyword research, metadata and content optimisation for organic visibility.

SEO
Content
Analytics

Dec 2025 — Feb 2026Remote

Front End Developer Intern

Gocourse Pvt Ltd

Built the front end of the Gocourse books website as a responsive, component-based UI.

React
HTML
CSS

03 Selected work Real screens from real builds

Things I’ve shipped.

Winner — Smart Ability Hackathon 2026

01 — Featured

EchoSense AI

Tinnitus assessment, sound therapy and doctor consultation on one platform, now being deployed for NIEPMD, Govt. of India.

Assessment
Prediction
Sound therapy
AI companion
Doctor consultation
3D ear viewer
Clinician console

React
TypeScript
Vite
Zustand
Three.js
Web Audio API
i18next
Django REST
JWT
HL7 FHIR R4
scikit-learn
SciPy
Playwright
Render

echosense/

EchoSense clinical report: audiogram with the matched tinnitus pitch, composite index and spectral signature

EchoSense sound therapy player generating notched broadband noise with a live output spectrum

EchoSense clinician caseload ranking patients who need attention

EchoSense clinician view with Shapley model attribution and a 3D cochlea coloured by the patient's thresholds

Keep scrolling

02Smart India Hackathon 2025

TwinCode

NAMASTE ⇄ ICD-11 dual-coding FHIR microservice

An MVP FHIR R4 service that dual-codes India’s traditional-medicine terminologies with WHO ICD-11 so EMR records stay interoperable. I built the data and mapping layer: Sentence-BERT/ClinicalBERT semantic mapping, spaCy + scikit-learn autocorrect and an engine that explains every assigned code.

NestJS
TypeScript
PostgreSQL
Next.js
FHIR R4
OAuth 2.0 / OIDC
Sentence-BERT

pramana/

Pramana IP and regulatory roadmap with a decision map and action plan

Pramana evidence assistant answering with citations and a confidence rationale

03Smart India Hackathon 2026

Pramana

AI-assisted IP & regulatory guidance for AYUSH innovators

A guided path from product classification to legal evidence, compliance, protection strategy and an IP-to-market roadmap. Hybrid RAG runs over official corpora, a verification engine citation-checks and NLI-validates every AI claim, and uncertain cases go to human IP facilitators.

Next.js 15
FastAPI
pgvector
BGE-M3
DeBERTa-v3 NLI
Bhashini
FHIR R4

sevps/operations

SEVPS control room: live map of ambulances, routes and signals held for green corridors

SEVPS driver navigation with the route to hospital and signals ahead scheduled for a green corridor

SEVPS paramedic flow recommending the best-prepared hospital

SEVPS hospital readiness board with beds, ICU, ventilators and an inbound ambulance alert

04Finalist — CIH ’26

SEVPS

Smart Emergency Vehicle Priority System

A real-time platform that gives ambulances signal priority, routes them over PostGIS spatial data with NetworkX and tracks live hospital readiness. Admin, driver, paramedic and hospital dashboards update live over WebSockets, and the estimators come with SHAP explanations.

Django Channels
Redis
PostGIS
NetworkX
YOLOv8
React
Docker

shn/injection

SHN fault-injection console: healthy React Flow topology with live telemetry and animated routes

SHN injecting a link failure between Router-03 and Router-04

SHN detects the failed link, diagnoses it with 90% confidence and starts rerouting

SHN operator portal: traffic rerouted through Router-05 with the live healing steps verified

SHN incident log recognising a repeat failure, with evidence, old and new routes and before/after metrics

05Network automation

SHN

Self-Healing Network

An observe → detect → diagnose → heal → summarise loop for network failures. Celery Beat drives a 2 s tick, NetworkX reroutes along shortest paths and telemetry streams to a React Flow topology. Incidents are explained by templates or by any OpenAI-compatible LLM, with automatic fallback.

React Flow
Django Channels
Celery
Redis
NetworkX
PostgreSQL

click/home

CLICK microlearning lesson explaining the three parts of a for loop

CLICK chapter test: fill in the missing parts of a C program in the code editor

CLICK chapter test complete with a celebrating DADA and XP added

06Live in my department

CLICK

C programming learning + practice platform

Built for the first-year students of my department: microlearning lessons, practice problems, tests and progress tracking. It ships as a PWA, a Windows installer, an Android APK and a VS Code extension, running on a hand-rolled, framework-free SPA on Cloudflare Workers with a Supabase Edge Function backend, load-tested with k6.

Vanilla JS
Cloudflare Workers
Supabase
Electron
VS Code API
k6
Playwright

‹ A

Academic WDBonline

Today

hi9:41

👋 Welcome to Academic WDB. Send your department, semester and subject.9:41

AI&DS · sem 4 · DBMS9:42

PDFDBMS_Unit_1-5_Notes.pdf18 pages · 2.4 MB9:42

▶Normalisation in DBMSyoutube.com · lecture video9:42

🔗 3 reference links for DBMS, sem 4. Good luck! 🍀9:42

Message

07WhatsApp Cloud API

Academic WDB

WhatsApp-based academic resource bot

Students message a number and get notes, videos and reference links for any subject, semester or department within seconds. Intent matching is regex-first with a Groq Llama-3.1 fallback, backed by a JWT-secured DRF API, Supabase and a React admin panel on Render.

Django REST
SimpleJWT
React
Supabase
Groq LLM
Docker
pytest

psychai/

PsychologyAI home: welcome back, latest wellness score and a Start PHQ-9 button

PHQ-9 questionnaire mid-way with answers selected

Assessment result: score 11, Moderate, confirmed by the AI prediction

AI wellness companion chat with supportive replies

Consultation request form with a preferred date and time

08Paper in preparation

PsychologyAI

AI/ML mental-health companion

PHQ-9 and BDI questionnaires, rehabilitation guidance, doctor consultation and an LLM-powered companion. I trained a scikit-learn classifier on PHQ-9 responses to estimate depression severity, and that work is being written up for publication.

Django
scikit-learn
pandas
Groq LLM
Bootstrap 5
Render
Vercel

More in the forge Everything else lives on GitHub. github.com/
darshansathish2006

04 Achievements Trophies, rounds cleared, rooms presented to

Hall of Flame.

Winner

Smart Ability Hackathon 2026

Organised by Rajalakshmi Engineering College with NIEPMD. Won with EchoSense AI, which is now being deployed at NIEPMD by the Govt. of India.

2× Winner

Smart India Hackathon — Internal Round

24-hour hackathon by the Govt. of India. Built TwinCode (SIH ’25) and Pramana (SIH ’26).

Winner

Kuttykoncepts Flagship Hackathon 2026

Organised by Rajalakshmi Engineering College, for generative AI and AI video generation.

Finalist

CIH ’26 · Rathinam Techzone

Cleared every round to reach the finals of the 24-hour hackathon, where SEVPS was built.

2 of 3 rounds

Adobe University Hackathon 2026

Cleared two of three rounds among 3,00,000+ participants nationwide.

AI NEXUS 2026

Presented to the Minister for AI, IT & Digital Services

Showed my internship work at the AI Innovators Expo of REC’s AI Tech Summit to Dr. Kumar R, Hon’ble Minister for Artificial Intelligence, IT & Digital Services, Government of Tamil Nadu.

4.75 / 5

FH Aachen · CyberSol

Graded 95% in the CyberSol International Summer Internship at the MASCOR Institute.

05 Speaking & leadership Rooms, mics and a podcast

Off the keyboard, on the mic.

Founder & Host

The student-led podcast

I conceived, organised and hosted a two-episode series with leaders from academia and industry, and handled everything from guest outreach and scripting to moderation and production.

Episode 01

with Mr. Marcel Stüttgen

Student life, computer science, the field of AI, and the skills Germany looks for in young engineers.

Full episode link coming soon

Episode 02

with Mr. Raj Vangapandu

Leadership, entrepreneurship, building a company from the ground up, life in Germany, and the motivation to keep going.

Full episode link coming soon

Product launch · Phase 1 & 2

CLICK deployment

Rolled CLICK out to the first-year students of my department in two phases, walking each batch through it live, from signing in to writing their first programs.

Guest speaker

Hackathon Awareness Session 2026

Invited by the Department of AI & DS at REC to coach juniors on hackathon strategy, team formation and problem selection.

Lead organiser & sole speaker

TechSavvy 2026 · MCC

Planned the curriculum and delivered every session of a Python problem-solving and HTML/CSS workshop for Class XII students.

Presenter

AI NEXUS 2026 · AI Innovators Expo

Presented my internship work to the Hon’ble Minister for AI, IT & Digital Services, Government of Tamil Nadu.

06 Publications & research Digital forensics · ML for mental health

Reconstructing what was deleted.

Under review · decision Oct 2026 AIFB 2026 · 2nd International Conference on AI for Forensics and Biometrics · IIT Dharwad

Ghost Evidence: An Explainable Framework for Digital Evidence Correlation and Event Reconstruction from Missing Primary Evidence

Andryan Durai Jebin, Darshan C S, Akshaya P, Dinesh G · Rajalakshmi Engineering College

When the primary artifact has been deleted or overwritten, Ghost Evidence reconstructs the probable digital event and its time interval from the secondary traces that survive. It stacks a forensic rule layer, temporal reasoning, interpretable ML correlation and Bayesian-style evidence aggregation.

Formal problem definition
Evidence-source taxonomy with reliability weighting
Expected-trace-absence feature
Controlled evidence-removal evaluation
Forensic-validity & governance analysis

In preparationPsychologyAI

Depression-Severity Classification from PHQ-9 Responses

A manuscript on a scikit-learn text classifier for depression-severity analysis, aimed at early-intervention outreach to at-risk users.

07 Skills Front end to model weights

The arsenal.

Languages: Python, C, C++, Java, JavaScript, TypeScript, SQL
Frontend: React, HTML5, CSS3, Bootstrap, Three.js, Web Audio API
Backend & APIs: Django, Django REST Framework, REST API design, JWT, Django Channels (WebSockets), Meta WhatsApp Cloud API
Databases: PostgreSQL, PostGIS, MySQL, MongoDB, SQLite, Supabase (Postgres, Storage)
AI / ML: Machine Learning, NLP, LLM integration (Groq, OpenAI-compatible APIs), Prompt Engineering, GenAI, NumPy, pandas, Neural Networks, Reinforcement Learning
DevOps & tools: Docker, Git, GitHub, Render, Vercel, Cloudflare Workers, VS Code Extension API, Playwright, k6