Dushyant Engineer · Builder · Creator
About Projects Contact

We build systems.

We build futures.

Your vision,
engineered.

From orbital defense systems to AI desktops — every project is a mission. Every line of code, a purpose.

Scroll down

To explore the work

Jet Jet Blueprint
Build in
Excellence

Excellence
that moves with you

Dushyant
Full-Stack Engineer

Every system is engineered from first principles — scalable architecture, real-time data pipelines, and interfaces that feel alive.

Dushyant
Full-Stack
Engineer
University Sharda University, Greater Noida
Degree B.Tech CS · AI & DS · 2025–2029
Backend Python · FastAPI · Flask · Node.js · Rust
Frontend React · Next.js · TypeScript · Tailwind · Three.js
AI / ML Scikit-learn · PyTorch · SHAP · PINN
Data & Infra PostgreSQL · Redis · Firebase · Docker · Git
Systems Tauri · Rust · IoT · ESP32 · WebSockets
Deployment Vercel · Render · Netlify · Railway
Version Control Git · GitHub · CI/CD

AI & Systems
Engineer

Precision Engineering at Scale

B.Tech Computer Science (AI & Data Science) at Sharda University, batch 2025–2029. Building production-grade systems across orbital defense, healthcare AI, fleet intelligence, and e-commerce.

AI-Driven Intelligence

From dyslexia detection via behavioral ML to cardiovascular risk prediction with SHAP explainability — embedding intelligence deep into every layer of the stack.

End-to-End Ownership

34,000+ orbital objects tracked. Sub-second WebSocket telemetry. IoT hardware to cloud inference. Every project owned fully, shipped completely.

Selected Work

Subtitle / Projects

Build in
Excellence
01

Vyoma

FastAPIWebSocket3D VizSSA
Vyoma

Autonomous Orbital Defense System

Enterprise-grade AODS with a multi-tenant FastAPI backend and immersive 3D orbit visualization. Persistent background schedulers ingest live TLE datasets and synthetic rocket telemetry, streamed via low-latency WebSockets to a Three.js client. Tracks 34,000+ orbital objects through a full SSA pipeline including trajectory propagation, collision probability analysis, CDM processing, debris risk field estimation, and Physics-Informed Neural Networks (PINN) for orbit correction. Autonomous evasion-maneuver engine predicts collision events 72 hours in advance and computes fuel-optimal avoidance maneuvers.

BackendFastAPI · Python · SQLAlchemy · Redis
FrontendThree.js · WebSockets · React
AI/MLPINN · Collision Prediction · Evasion Engine
DataTLE Ingestion · CDM Processing · Digital Twins
Objects Tracked34,000+
Prediction Window72 hours ahead
Not on GitHub

AI Cognitive Assessment & Training Platform

End-to-end platform detecting dyslexia and dysgraphia via behavioral analysis. Next.js + Tailwind + shadcn UI frontend captures structured stroke data ({x, y, time, pressure, type}) from a digital writing canvas and reading modules, sending it to a FastAPI backend. The backend pipeline extracts features — speed, stroke consistency, pressure variation, pauses — feeds them into an ML classifier with SHAP-based explainability, and generates adaptive training plans with XP tracking. Redis handles rate limiting and dashboard caching; PostgreSQL via SQLAlchemy stores users, assessments, results, and tasks. Recharts powers visualization. Frontend on Vercel, backend on Render.

FrontendNext.js · Tailwind · shadcn UI · Recharts
BackendFastAPI · SQLAlchemy · PostgreSQL · Redis
AuthJWT · Google OAuth
MLRandom Forest · SHAP Explainability
FeaturesAdaptive Training · XP System · Progress Tracking
DeployVercel (FE) · Render (BE)
View on GitHub
Mentis
02

Mentis

MLNext.jsFastAPICanvas
03

Velos

IoTLeaflet.jsChart.jsFleet
Velos

Fleet Health & Risk Intelligence Platform

Full-stack IoT fleet intelligence platform with a dark-themed Vanilla JS frontend (Chart.js analytics, Leaflet.js GPS tracking), FastAPI async backend, and PostgreSQL. JWT auth with role-based access. Health engine normalizes sensor values (RPM, engine temp, battery, tire pressure, speed) using weighted degradation logic to compute 0–100 health scores. Risk classifier categorizes vehicles into Low/Moderate/High/Critical states. Computes Fleet Volatility Score (std dev across health scores) and Fleet Risk Exposure Index. IoT-ready: ESP32 + NEO-6M GPS + SIM7600 GSM hardware integration path for real-time OBD-II CAN bus telemetry.

FrontendHTML · CSS · Vanilla JS · Chart.js · Leaflet.js
BackendFastAPI · Pydantic · PostgreSQL
AuthJWT · Role-Based Access Control
IoT HardwareESP32 · NEO-6M GPS · SIM7600 GSM · OBD-II
IntelligenceHealth Scoring · Risk Classification · Predictive Life
AnalyticsFleet Volatility · Risk Exposure Index · Rankings
View on GitHub

Cardiovascular Monitoring & Prediction Ecosystem

Hospital-centric platform where hospitals register via Firebase Auth, manage patients in Firestore, and submit cardiovascular parameters to a Flask /predict endpoint. Random Forest model computes heart disease probability; SHAP-based explanation engine highlights contributing factors. Longitudinal timeline enables trend analysis. What-if simulation engine lets doctors model health improvements in real time. IoT upgrade path for continuous remote cardiovascular monitoring.

Auth & DBFirebase Auth · Firestore
BackendFlask · Python · Random Forest · SHAP
FrontendReact · Timeline UI · Charts
FeaturesWhat-If Simulator · Trend Analysis · CSV Export
IoT HardwareESP32 · AD8232 ECG · MAX30102 · GSM
IntelligenceRisk Prediction · Explainable AI · Anomaly Detection
View on GitHub
VitaPulse
04

VitaPulse

FirebaseFlaskMLIoT
05

Gaurk

ReactFastAPIViteLive
Gaurk

Production E-Commerce Platform

Full-stack e-commerce from scratch: React + Vite storefront and admin panel, FastAPI backend, PostgreSQL via SQLAlchemy ORM. Full purchase lifecycle: browse → variant select → cart → checkout → Razorpay payment → confirmation. Admin panel handles variant-level inventory, order management, promo codes, newsletters. Firebase for auth. Shiprocket for shipping automation with live tracking.

FrontendReact · Vite · Tailwind
BackendFastAPI · SQLAlchemy · PostgreSQL
PaymentsRazorpay Payment Gateway
ShippingShiprocket API · Live Tracking
Auth & NotifyFirebase Auth · Push Notifications
DeployRender · Netlify · Git
Live atgaurk.shop

Desktop AI Productivity Environment

Privacy-first, local-first desktop AI ecosystem built with Tauri v2 + Rust backend and React + TypeScript frontend. Modular Rust agent engine with typed Tauri IPC commands exposes PLAN, BUILD, TEST, DEBUG, and SUMMARIZE agent modes. Real-time event emission. Zustand for state, CodeMirror for editing. Memory persistence and autonomous operation from planning to final output.

RuntimeTauri v2 · Rust
FrontendReact · TypeScript · Zustand
EditorCodeMirror
Agent ModesPLAN · BUILD · TEST · DEBUG · SUMMARIZE
IPCTyped Tauri IPC · Event Emission System
ArchitectureLocal-First · Privacy-Focused · Memory Persistence
Not on GitHub
DUAI
06

DUAI

TauriRustReactAI
Projects shipped 6+
Based in India
Stack depth Full
Local time --:--
GLOBAL
Available for opportunities

Build anything,
from anywhere.

Open to global collaborations, freelance work, and full-time opportunities. Let's build something exceptional together.

Download CV

6+

Enterprise-grade systems built

From space to healthcare — each system built to production standards, with scalability and elegance at its core.

For
inquiries