Systems Programmer & AI Engineer
Second-year B.Tech CSE student at VIT Chennai — specialising in Systems Programming and AI, with a strong grasp of how computers work at a low level. My code is clean, modular, and performance-conscious.
CS student, builder, and systems thinker from VIT Chennai.
Hi, I'm Rajarshi Saha — a Computer Science student at VIT Chennai, currently in my second year of B.Tech. I specialize in Systems Programming and Artificial Intelligence, with a strong foundation in how computers actually work at a low level.
I've built everything from AI-powered web apps and VTOP automation tools to deep learning models that predict customer churn and stock prices. My code tends to be clean, modular, and performance-conscious.
I'm actively preparing for national hackathons and competitive technical roles. When I'm not coding, I'm studying algorithms, OS internals, and the math behind machine learning.
Discipline compounds.
Skill scales.
Consistency wins.
Computer Science
Chennai Campus
Projects Built
Technologies
From production-grade web apps to deep learning research projects.
Web App · AI · NLP
An AI-powered mental health companion providing empathetic, conversational support using NLP. Designed as a safe, anonymous space.
Conversational AI with empathetic generation
Context-aware dialogue management
Mood tracking & sentiment analysis
Web App · Automation
A streamlined web client for VIT's VTOP portal, delivering a faster, cleaner experience to access attendance, grades, and timetables.
Minimal alternative to official portal
Attendance & grade dashboard
Session-based secure login
Web App · Social Impact
A web platform built for an NGO in West Bengal to showcase their efforts in education and empowerment. Built during SIH.
Outreach-focused UI for donors
Responsive, mobile-first interface
Deployed to Vercel
Machine Learning · BI
An end-to-end ML pipeline predicting telecom customer churn, comparing multiple classifiers with rigorous feature engineering.
EDA & feature engineering pipeline
Logistic Regression, Random Forest & XGBoost
AUC-ROC evaluation
Deep Learning · Time Series
Deep learning model forecasting stock price trends using LSTM networks to learn long-range temporal dependencies.
60-day sliding window sequence
LSTM with forget/input gates
BPTT training
Deep Learning · CV
Custom CNN architectures for image classification, built from scratch with systematic data augmentation.
Conv → Pooling → ReLU architecture
Data augmentation via ImageDataGenerator
OpenCV preprocessing
Languages, frameworks, and tools I work with daily.
Where I'm pushing myself this year — tracked and in motion.
40% — 1 of 3 shipped
20% — Preparing for upcoming events
35% — Deep into OS & memory
10% — Researching active repositories
Building something interesting? Want to collaborate? Let's talk.
I'm open to internships, project collaborations, hackathon teams, and interesting technical conversations. The fastest way to reach me is email.