Devansh.
Engineer working across backend, ML and deep learning.
Final-year CSE student. I built a full platform for a coaching institute as a freelancer, interned at IIT BHU and ImpactBridge, and have two papers accepted at IEEE conferences this year.
I build across backend and ML, from payment flows and APIs to deep learning models small enough to run on a phone.
On the backend side that means payments, logins, exam engines and job queues. On the ML side it's deep learning research at IIT BHU and two papers accepted at IEEE conferences this year. Lately I've been getting into distributed systems, building queues, locks and rate limiters on Redis to understand how things scale.
Read full about →Where I've shipped
Built LightDep, a 3.96 MB distilled model for depression screening that runs on the device. The paper got accepted at IEEE ANTS 2026, and I'm joint first author.
Built the whole platform for a JEE/NEET coaching institute on my own: enrollment, Cashfree fee payments, an exam engine, and separate dashboards for five roles.
Built the volunteer side of ImpactBridge, a platform that connects NGOs with volunteers: signup and login, an apply flow with resume upload, and an email to the NGO for every application.
Production & systems work
Accepted papers
LightDep: Knowledge-Distilled Multimodal Depression Screening for Mobile and Edge Devices
Devansh Pandey*, Prateek Goyal*, Dr. Amit Vishwakarma1, Dr. Arijit Roy2, Dr. Om Jee Pandey3
1IIITDM Jabalpur·2IIT Patna·3IIT BHU
Structure-Aware Drug–Target Interaction Prediction via Pretrained Language Models and Cross-Attention Fusion
Srajal Tiwari, Dr. Dolly Sharma, Prateek Goyal, Devansh Pandey, Anamika Pal, Abhinav Bachchas
Still shipping.
My inbox is open.
Whether it's a role, a project, or a question about something I built, I'll reply.
When I'm not shipping: quizzing, Indian history rabbit holes, and the occasional build log on Hashnode.