Machine Learning • Quant Finance
Stock Market Direction Predictor for Indian Equities
A full-stack stock analysis platform built to support equity screening and short-term direction prediction using real-time
Indian market data.
- Built a hybrid pipeline with EMA, SuperTrend, moving averages, and a Random Forest model for next-day forecasting.
- Used walk-forward backtesting and precision-focused evaluation to assess prediction reliability.
- Integrated volatility-aware risk estimation to improve decision support.
- Designed an interactive interface with stock search and scalable backend integration.
NLP • Knowledge Graph • Rule-Based AI
Movie Savant: Knowledge-Graph-Based Movie Expert System
A hybrid movie expert system for natural-language querying over films, actors, directors, genres, and release years.
- Combined knowledge-graph ingestion, rule-based query handling, fuzzy title matching, and contextual memory.
- Built a CSV-to-database pipeline with entity–relationship modeling for stronger answer reliability.
- Designed a ChatGPT-style frontend with query suggestions, confidence-aware responses, and conversation history.
- Focused on efficient, scalable retrieval for multi-turn movie conversations.