8.38
Current CGPA
B.Tech Computer Science & Engineering, LPU.
Open to engineering internships
I build backend systems, and I document where each one stops.
Backend programming · API design · Database design
01 / THE RECORD
Four facts, stated plainly.
2028
Expected graduation
Third-year coursework across DSA, DBMS, OS and networks.
4
Documented projects
Each one written up with its stack and its limitations.
2
Open-source platforms
Contributor on AdityaNet and ASTRA research codebases.
Nothing here is inflated. Where a project is a mockup, it says so; where a contribution is partial, it says that too.
02 / THE ENGINEER
I want to know what a system is doing underneath.
I am a B.Tech Computer Science & Engineering student at Lovely Professional University, focused on backend development and software engineering. My work sits around three things: backend programming, API design, and database design. I write most of it in Python and JavaScript.
The projects on this site were built to answer questions rather than to fill a resume. A commerce platform for a jeweller in Kanpur that real customers use, where the auth, the database schema and the deployment were all mine to get right. A face-recognition logger to understand how image classification actually gets trained and served. Two research platforms where I contributed to real verification and rendering code alongside people doing the science.
I am looking for a Software Engineering, Backend Engineering, or Full-Stack Engineering internship where I can build reliable API endpoints and database-backed services, and learn distributed-systems work from engineers who have shipped it.
Currently
B.Tech CSE, Lovely Professional University
Seeking
Backend / software / full-stack engineering internships
Location
Kanpur, Uttar Pradesh, India
03 / OPERATING RULES
Four rules the work is built under.
Not values — habits. Each one shows up in the case write-ups on the work page.
R-01
Build it from scratch first.
Local scripts and hand-written layouts before frameworks. The commerce storefront ships without React or a build pipeline; the attendance logger does its own frame capture, preprocessing and classification. I want to see how the pieces connect before I let a library hide any of it.
R-02
Verify against real output, not intent.
Endpoints get called, classifier output gets checked against test data, and type configurations get tested rather than assumed. Code that has not been run against something is not finished.
R-03
State the limitation before someone finds it.
Every project on this site ends with what it cannot do. The attendance system can be spoofed with a photograph. The commerce engine has no cart or checkout yet, and its README says so before it says anything else. Stating it is cheaper than having it discovered in an interview.
Seen in
Every case ↗
R-04
Leave the repository readable.
Readable code, documented constraints, and directories organised so the next person can find things. A project nobody else can run is a project that only ever existed on my machine.
Seen in
github.com/Lakshya172 ↗
04 / SELECTED WORK
Four projects, examined.
Problem, what I built, the stack, what I learned, and the limitation each one still carries. Full write-ups on the work page.
Rinky Commerce Engine
A live commerce platform for a jeweller in Kanpur. Node.js and Express REST API, MongoDB Atlas, session auth with role-based admin access, wishlists, and Brevo transactional email — deployed on Render and serving real customers.
Node.js
Express
MongoDB
Mongoose
Facial Attendance System
A local face-recognition check-in logger. OpenCV Haar Cascades detect and crop faces, a scikit-learn KNN classifier identifies them, and check-ins land in CSV — with a Streamlit dashboard reading the log live.
Python
OpenCV
Scikit-learn
Streamlit
AdityaNet
Contributor to a verifiable solar-flare research platform for Aditya-L1 satellite data — Astro/TypeScript interfaces, automated verification budgets, and reproducibility documentation for Level-1 dataset ingestion.
Python
Astro
Three.js
TypeScript
ASTRA
Contributor to a stellar transient classification platform — server-side ONNX Runtime inference, interactive 3D WebGL space mapping with Three.js, and dense time-series charts for stellar metrics.
Next.js
Three.js
ONNX Runtime
TypeScript
05 / CONTACT
Open channel.
Available for backend, software, and full-stack engineering internships. The fastest route is email.