Computer Engineering student at Thapar Institute (batch of 2027) who builds web apps, ML models, and tools that use LLMs. I'm currently an SDE intern at Vardhaman Yarn and Threads Ltd. (VYTL), where I built and deployed two internal PHP/MySQL apps. I like full-stack development, prompt design, applied NLP, and edge devices. Away from the keyboard I'm drawn to business strategy, customer research, event management, and stage performance, and I'm happiest where tech meets people.
- Full Stack: PHP/MySQL backends, React.js and Next.js with TypeScript on the front, REST APIs, and Firebase-backed apps.
- AI/ML: Python models for text classification and prediction, plus Claude and the NVIDIA Gen AI stack for classification and automation tasks.
- Prompt Engineering: Writing and testing prompts in Claude, Gemini, ChatGPT, and Cursor. In the Digital VYTL Dashboard, a Claude planning assistant sits behind a server-side proxy, so the API key never reaches the browser.
- Business & Strategy: Customer surveys, go-to-market plans, and financial models, such as the Omni Solutions venture below.
- Systems & Embedded: Arduino, MATLAB, and RoboStudio projects, packet-level networking in C++, and requirements work written up as SRS documents with UML diagrams.
- SDE Intern @ Vardhaman Yarn and Threads Ltd. (VYTL)
- Visitor Management System: replaced paper gate registers with a PHP/MySQL system for 100+ daily visitors. QR passes are signed with HMAC and verified on the server. Check-in time dropped by 70% and manual data entry by 80%.
- Digital VYTL Dashboard: task board and intranet portal in plain PHP and vanilla JS, with an AI planning assistant that calls Claude through a server-side proxy.
- Open to collaborate on: ML/LLM projects, full-stack apps, open-source tools
- Learning: NLP, system design, SRS writing, and consulting-tech, plus how product and business teams make decisions
| Project | What it does | Stack |
|---|---|---|
| Fake Job Posting Detector | Flags fraudulent listings across 17,880 postings using TF-IDF features and 60+ fraud rules; Naive Bayes scored 97.5% accuracy. Flask API scrapes and checks a posting straight from its URL. | Python, Flask, scikit-learn, NLTK |
| DPI Engine | Reads SNI from encrypted HTTPS traffic to name 18+ apps, with a Random Forest behind a FastAPI service and a CLI that blocks by app, domain, or IP. Forked and heavily modified. | C++, Python, FastAPI, Scapy |
| AMR Surveillance, India (live) | Maps antibiotic resistance across 28 states from 3 datasets and 1M+ records; a 94.6%-accurate model suggests antibiotics from patient history. | Python, Pandas, Streamlit, Folium |
| Ridezon | Ride-sharing web app for verified students: create and join rides, chat, split costs. Four-person team; IEEE-style SRS with UML and DFDs. | React.js, TypeScript, Node.js |
| Stock Price Prediction | Pulls history from yfinance and forecasts the next 30 days with a sliding-window regression, scored by train/test RMSE. | Python, scikit-learn, yfinance |
| E-Commerce Funnel Analytics | CTE-based SQL over a five-stage funnel for conversion, drop-off, and AOV, shown in a self-serve dashboard. | SQL (BigQuery), Looker Studio |
| Omni Solutions | Venture study for an elderly fall-detection product: 184 survey responses, go-to-market plan, and a 3-year financial model. | Market research, financial modeling |
Also: co-built the Virsa Society website (Next.js, TypeScript, Tailwind CSS) with 20+ commits. Full list in my Portfolio.
- Business thinking: Omni Solutions pushed me out of the editor and into interviews, spreadsheets, and pricing. The survey showed 41.3% of respondents worried about privacy and 40.2% about false alarms, and both findings shaped the product.
- Virsa Society (TIET): Technical, Publicity and Outreach Lead. I perform Bhangra, anchor events, lead teams, and help juniors find their feet.
- Mudra Society (TIET): Core Manager and Executive Board member, running backstage operations at large events.
- Earlier: Silver medal in inter-school swimming, school band leader, and a French language award.
- NVIDIA: Getting Started with Deep Learning, Generative AI with Diffusion Models, Jetson Nano Deployment, Accelerated Computing
- Training: Robo-AI Industrial Training (45 days), Android App Development (Kotlin, Jetpack Compose, Firebase), Frontend Development at Vidya Corp, Data Analyst Bootcamp at Alex Analytics
