Projects

Customer-discovery-led edtech concept that pivoted to AI-guided personalized learning after 124 user interviews.

A Streamlit investment dashboard that backtests step-up SIP strategies, compares portfolio XIRR against benchmarks, and cleans Yahoo Finance data before surfacing risk, drawdown, and rolling-return analytics.

A Streamlit market-seasonality app that identifies which weekday historically delivers the lowest closing prices and pairs that signal with rank, return, volatility, and win-rate analytics.

A dual-market breadth-analysis tool that tracks how many stocks in US and Indian indices trade above key moving averages to surface overbought and oversold sentiment shifts.

An ongoing football analytics project that uses age, position, percentile-based peer ranking, and machine learning to estimate whether elite players can sustain peak performance deeper into their thirties.

Full-stack Human-AI research platform studying how LLM commentary changes trust, risk perception, and decision quality in retail investing.

An OMSCS AI Ethics and Society project on employment bias mitigation using intersection-aware reweighting and measurable fairness diagnostics.

A privacy-first local stock-research assistant that combines multi-agent orchestration, RAG over financial documents, Ollama-hosted LLMs, and a Dockerized Streamlit interface.

OMSCS trading-systems project exploring reinforcement learning and decision-tree strategies for algorithmic trading.

Georgia Tech HCI redesign focused on improving ticket-booking usability, clarity, and flow under high-demand conditions.

A human-centered volunteering platform prototype that helps NGOs match with volunteers while simplifying permissions, funding, and event logistics through organizer-focused dashboards.

Published applied-ML project using Fantasy Premier League data to recommend players beyond simple ROI and fan favoritism bias.

Frontend-focused academic website template for research labs to showcase publications, projects, and team members more maintainably.

Systems prototype that monitors VM CPU load and boots standby instances using Python, libvirt, and QEMU/KVM.

A text simplification project aimed at making text easier to read while retaining original meaning, with accessibility-oriented motivation.

A C++ information-security project implementing RSA, ElGamal, ECC, and SHA-1-backed digital signature workflows with NTL big-integer arithmetic.

NLP comparison of VADER and BERT on EPL Twitter data to measure fan sentiment, semantic nuance, and social-media polarity.

Ruby on Rails MVC and JSON API project for managing a relational book catalog with clean CRUD workflows.

Security-focused computer vision system using MTCNN, ResNet-50 embeddings, and threshold tuning to eliminate false positives.

Responsive PHP/MySQL learning platform with secure authentication, AJAX filtering, progress tracking, and admin tooling.

Remote-sensing and modeling project quantifying urban warming, vegetation loss, and temperature anomalies in Bangalore.

Full-stack campus event platform that centralizes discovery, registration, and scheduling across NIT Calicut organizations.

A 100% Python automation script that extracts housing society billing data from PDFs and turns it into structured Excel summaries for faster local administration.

A live multiplayer mystery game for Ragam 2020 that combined a Django investigation portal with Discord automation to run a 585-registration online event.