RoshanRaj M
I design, build and ship production systems — end to end. Full-stack systems and applied AI/ML — from first commit to production deploy.
- 6+
- Products shipped
- 3
- Industry internships
- 4
- Production repos
- AI · FS
- Dual focus
An engineer who finishes.
I'm a final-year B.Tech (CSE — AI & ML) engineer at Jain University who builds complete products, not prototypes.
- Focus
- Full-Stack + Applied AI/ML
- Based in
- Bangalore, India
- Degree
- B.Tech CSE (AI & ML)
- Open to
- Engineering roles & freelance
I work across the full stack — React & Next.js on the front, FastAPI & Python on the back, Supabase/PostgreSQL for data — and ship them with Docker, a Caddy reverse proxy and auto-deploy pipelines. My flagship, Udaya LMS, runs live in production with JWT auth, an AI mentorship engine and WhatsApp integration.
My foundation is AI/ML. Across internships at Tech Mahindra and Six Phrase I built real computer-vision systems — safety-violation detection and real-time YOLOv8 vehicle analytics — so I'm just as comfortable training a model as I am wiring up the deployment around it.
I care about the whole lifecycle: clean architecture, secure auth, reproducible deploys and an interface that feels effortless. That bias toward finished, deployed software is what I bring to a team.
Education
B.Tech — CSE (AI & ML)
2022 — 2026Jain (Deemed-to-be University), Bangalore
CGPA 7.4 · Pursuing
Higher Secondary Education
2019 — 2021Seva Mandir Post Basic Higher Secondary School, Kerala
94.5%
Where I've worked
Six Phrase
AI & Machine Learning Intern
Jan 2025 — Apr 2025
- Built a real-time vehicle detection system with YOLOv8 and OpenCV.
- Shipped a Flask web app for live traffic analysis with dashboards and alerts.
Tech Mahindra
AI Intern
Dec 2024 — Mar 2025
- Developed a computer-vision system to detect safety violations in images.
- Worked on AI model integration and real-time image processing.
Miles Education
Machine Learning Intern
Apr 2024 — Jun 2024
- Cleaned and analyzed healthcare data with Python and SQL.
- Capstone: combined datasets to generate actionable insights.
A toolkit for shipping.
Comfortable across the stack — from training computer-vision models to provisioning the reverse proxy that serves them.
Languages
Day-to-dayFrontend
InterfacesBackend
APIs & servicesData & DB
PersistenceAI / ML
AppliedCloud & DevOps
Ship itProducts, not prototypes.
Real systems, designed and deployed end to end. Two flagship builds lead the way.
Udaya LMS
Learning Management System for tuition centers
A production LMS with separate student and teacher portals, an AI mentorship engine and WhatsApp notifications — deployed end to end behind a Caddy reverse proxy.
- Role-based dual portals (students + teachers) secured with JWT auth
- Assignments engine: create with attachments, submit, and grade
- AI mentorship generates personalized report cards & insights (Gemini)
- WhatsApp notifications via Evolution API integration
- Dockerized with a Caddy reverse proxy and auto-deploy scripts
2
Role-based portals
Live
On Vercel
AI
Mentorship engine
NJ Quotation System
Quotation & invoicing tool, shipped as a Windows app
A full-stack quotation system for NJ India, packaged into a one-click Windows installer so non-technical staff can run the whole stack locally.
- React + Vite frontend over a Python backend API
- End-to-end client quotation generation workflow
- Packaged as a native Windows installer with Inno Setup + C# bootstrapper
Olrac Signage
Cloud digital-signage platform & ad controller
A cloud signage backend that schedules and distributes advertising content to display devices, with web and Android TV players fetching playlists.
- FastAPI backend with async SQLAlchemy + asyncpg over Supabase
- Playlist resolution & scheduling for distributed displays
- Background tasks track offline devices in real time
Vision Assistant
Scene-to-speech aid for the visually impaired
Captures webcam input, generates image captions, summarizes them and speaks the result aloud to assist visually impaired users.
Traffic Vehicle Detection
Real-time YOLOv8 traffic analytics
Detects and counts vehicles from live video feeds and surfaces congestion analytics through a Flask dashboard with alerts.
Railway Management System
Booking & scheduling with live DB ops
Manages ticket booking, train schedules and passenger records with real-time database operations that cut manual errors.
I own the whole stack.
Not just the code — the way it ships. Containers, reverse proxies, managed databases and the integrations that make a product real.
Udaya LMS — deployment topology
React + Vite
Vercel
Caddy
Reverse proxy · TLS
FastAPI
Dockerized backend
Supabase
PostgreSQL
Integrations wired into the API
Containerized deploys
Multi-service apps packaged with Docker & Docker Compose for reproducible environments from laptop to production.
Reverse proxy & TLS
Caddy fronting FastAPI services and static frontends, with automatic HTTPS and clean routing.
Auth & security
JWT-based authentication with role-aware access control across student/teacher and admin surfaces.
Managed Postgres
Supabase / PostgreSQL with async SQLAlchemy + asyncpg for performant, type-safe data access.
CI & auto-deploy
Shell auto-deploy scripts and Vercel pipelines that turn a push into a live release.
Desktop packaging
Full stacks bundled into one-click Windows installers with Inno Setup for non-technical end users.
Third-party integrations
Wiring real services in: WhatsApp via Evolution API, Gemini AI, and Android TV clients.
Live from GitHub.
Pulled in real time from the GitHub API — the repositories and contribution graph below update themselves.
—
Repositories
0
Stars
—
Followers
—
Languages
Contribution graph
Let's build something.
Have a role, a project, or just want to talk shop? My inbox is open.
Based in Bangalore, India — working with teams everywhere.