CS student. Building at the intersection of AI and systems.
Computer Science undergraduate at Nirma University focused on backend systems, scalable APIs, and AI-powered automation. Experienced in building full-stack applications with secure authentication, real-time communication, and LLM-integrated workflows.
Real-time voice ordering agent using Twilio Media Streams with a dual STT pipeline for low-latency telephony conversations.
Engineered a multilingual NLP pipeline with custom VAD and interrupt handling. Built a Gemini-powered RAG system with embeddings for conversational menu search. Implemented a revenue intelligence engine for upsell prioritization and combo recommendations using PostgreSQL.
Full-stack microservices POS platform with REST APIs, 26 models, JWT-based RBAC, and real-time kitchen sync.
Engineered real-time kitchen sync via Socket.IO and inventory tracking using Prisma and PostgreSQL. Integrated FastAPI AI service for demand forecasting, waste scoring, and inventory simulation.
AI-powered financial analytics platform automating equity research across 500+ stock datasets.
Deployed a multi-agent system to generate earnings summaries and sector insights. Delivered visual dashboards with actionable investment insights using real-time market data.
End-to-end AI pipeline that extracts, summarizes, and transforms research papers into podcasts, presentations, and educational videos.
Extracts text from PDFs (including scanned documents via Tesseract OCR) and uses Google Gemini to generate plain-language summaries. Produces podcast scripts enriched with analogies, converts them to audio via gTTS, and overlays audio on video clips with MoviePy. Automatically builds themed PowerPoint presentations from research content.
Streamlit app that researches, generates, and refines full blog posts using Groq LLMs with real-time insights, keyword suggestions, and markdown export.
Combines a research assistant, dynamic title and keyword generator, full blog writer, and FAQ builder into a single Streamlit workflow. Powered by LangChain and Groq LLM API (LLaMA 3 70B, LLaMA 3.3 70B, Gemma 7B). Includes an AI chat assistant for iterative refinement and one-click markdown downloads for both research notes and final output.
A federated learning framework enforced with quantum mechanisms for detecting financial fraud across distributed systems while preserving data privacy. Presented at ACM MobiHoc 2025, Houston, USA.
A quantum-secured framework combining explainable AI and machine unlearning for phishing detection in IoT environments. Presented at IEEE GLOBECOM 2025 Workshops.
Whether it's a project, opportunity, or just a hello.